thin zinc oxide and cuprous oxide films for photovoltaic
TRANSCRIPT
Thin zinc oxide and cuprous oxide films for photovoltaic applications
A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL
OF THE UNIVERSITY OF MINNESOTA BY
SeongHo Jeong
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
Eray S. Aydil, Advisor
August 2010
Copyright © 2010 by SeongHo Jeong
i
Acknowledgements
I give heartfelt thanks to all the people who gave me a great support during my
time in Minnesota. I know it cannot be sufficient to express my thanks to them with
words, but I believe they know my mind.
First of all, I am deeply thankful to my advisor Eray Aydil. He not only provided
me a perfect support for my study but also showed me an endless passion for research
and life. I am very proud of being his student.
I also would like to thank my wonderful colleagues, Mike, Janice, Kurtis, Emil,
Will, Bin, Ankur, Selin, Boris, Neema, Melissa and Sriharsha. They always were willing
to give me a hand and all the discussion with them was joyful and stimulated me. I cannot
forget the help from the CEMS Korean students. They always supported me in and out of
the school and gave me courage and support. The life in school was warm and cheerful
because they were with me. I wish good luck to them all in their research.
I am also grateful to my friends and company colleagues, Inkyo, Yeonwoo,
Seokho, Seokheon, Sungho. Although they are far from me, they always pray for me and
show unchanged friendship. I am looking forward to seeing them all.
I have to express my appreciation to my family in Korea, parents, uncles, aunts,
brothers and sisters for their love and support during my study. I could not repay their
kindness.
ii
Finally, I would like to give my deepest thanks to my wife, Sungshin and
daughter, Seyoon. They are always behind me. This work would not be possible without
their support and understanding. Thank you and love you.
iii
Dedication
with love to my family
iv
Abstract
Metal oxide semiconductors and heterojunctions made from thin films of metal
oxide semiconductors have broad range of functional properties and high potential in
optical, electrical and magnetic devices such as light emitting diodes, spintronic devices
and solar cells. Among the oxide semiconductors, zinc oxide (ZnO) and cuprous oxide
(Cu2O) are attractive because they are inexpensive, abundant and nontoxic. As
synthesized ZnO is usually an intrinsic n - type semiconductor with wide band gap (3.4
eV) and can be used as the transparent conducting window layer in solar cells. As
synthesized Cu2O is usually a p - type semiconductor with a band gap of 2.17 eV and has
been considered as a potential material for the light absorbing layer in solar cells. I used
various techniques including metal organic chemical vapor deposition, magnetron
sputtering and atomic layer deposition to grow thin films of ZnO and Cu2O and
fabricated Cu2O/ZnO heterojunctions. I specifically investigated the optical and electrical
properties of Cu2O thin films deposited on ZnO by MOCVD and showed that Cu2O thin
films grow as single phase with [110] axis aligned perpendicular to the ZnO surface
which is (0001) plane and with in-plane rotational alignment due to (220)Cu2O || (0002)ZnO;
[001]Cu2O || [1210]ZnO epitaxy. Moreover, I fabricated solar cells based on these Cu2O/ZnO
heterojunctions and characterized them. Electrical characterization of these solar cells as
a function of temperature between 100 K and 300 K under illumination revealed that
interface recombination and tunneling at the interface are the factors that limit the solar
cell performance. To date solar cells based on Cu2O/ZnO heterojunctions had low open
v
circuit voltages (~ 0.3V) even though the expected value is around 1V. I achieved open
circuit voltages approaching 1V at low temperature (~ 100 K) and showed that if
interfacial recombination is reduced these cells can achieve their predicted potential.
vi
Table of Contents
Abstract........................................................................................................................ iv
List of Tables................................................................................................................ x
List of Figures............................................................................................................... xi
Chapter 1 Introduction............................................................................................. 1
1.1 Cu2O/ZnO heterojunction and solar cell application.............................. 1
1.2 Review of relevant ZnO and Cu2O properties........................................ 3
1.2.1 ZnO properties......................................................................... 3
1.2.2 Cu2O properties....................................................................... 5
1.3 Thesis organization................................................................................. 7
Chapter 2 Properties of ZnO thin films deposited by metal organic chemical
vapor deposition...................................................................................... 12
2.1 Introduction............................................................................................ 12
2.2 Metal organic chemical vapor deposition of ZnO thin films.................. 14
2.2.1 Chemical vapor deposition (CVD) process............................. 14
2.2.2 Chemical vapor deposition (CVD) system.............................. 16
2.2.3 ZnO film deposition kinetics................................................... 19
2.3 The dependence of ZnO film morphology on the substrates................. 22
2.4 Structural properties of ZnO thin films grown on r-plane sapphire....... 25
vii
2.4.1 Epitaxy between ZnO and r-plane sapphire............................ 26
2.4.2 Dependence of crystallinity on deposition temperature.......... 29
2.4.3 Surface morphology................................................................ 32
2.4.4 Two-step deposition................................................................ 35
2.5 Plasma enhanced chemical vapor deposition of ZnO............................. 36
2.5.1 Morphology of ZnO thin films deposited by PECVD............. 37
2.5.2 Effect of deposition pressure on the film morphology and
optical transmittance................................................................ 42
Chapter 3 Properties of ZnO thin films deposited by radio frequency
magnetron sputtering............................................................................... 48
3.1 Introduction............................................................................................ 48
3.2 Radio frequency magnetron sputtering for ZnO thin film...................... 49
3.3 The structure and optical properties of sputtered ZnO films.................. 51
3.4 Effect of post sputtering annealing on the structure and electrical
properties of ZnO thin films................................................................... 56
Chapter 4 Properties of Cu2O thin film deposited on ZnO by metal organic
chemical vapor deposition....................................................................... 63
4.1 Introduction............................................................................................ 63
4.2 The structure and electrical properties of Cu2O films on ZnO............... 64
4.2.1 Experiment............................................................................... 64
viii
4.2.2 Results and discussion............................................................ 67
4.2.3 Conclusion.............................................................................. 78
4.3 Heteroepitaxy of Cu2O thin film on ZnO.............................................. 78
Chapter 5 Cu2O/ZnO heterojunction and photovoltaic application......................... 87
5.1 Overview................................................................................................ 87
5.2 Cu2O/ZnO heterojunction solar cell fabrication.................................... 88
5.3 Solar cell characterization...................................................................... 90
5.3.1 Characterization background................................................... 90
5.3.2 Characterization equipment.................................................... 93
5.4 Photovoltaic characterization for Cu2O/ZnO solar cells........................ 96
5.5 Temperature dependent J-V characterization for Cu2O/ZnO solar cells 99
5.5.1 Background on the theory of temperature dependent solar
cell J-V characteristics............................................................ 99
5.5.2 Analysis of J-V data for Cu2O/ZnO solar cells....................... 103
5.5.2.1 Space charge limited J-V characteristics in ZnO in
the Dark.................................................................... 103
5.5.2.2 J-V characteristics under illumination....................... 108
5.5.3 Interpretation of the J-V-T analysis......................................... 112
5.5.3.1 Interfacial recombination........................................... 112
5.5.3.2 Tunneling................................................................... 114
5.5.3.3 Auger depth profile................................................... 119
ix
Chapter 6 Summary and future direction................................................................. 123
6.1 Summary................................................................................................. 123
6.2 Future direction....................................................................................... 126
Bibliography................................................................................................................. 130
x
List of Tables
Table 1.1. Fundamental properties of ZnO. Unless indicated otherwise, all
properties are at room temperature. “c” and “a” denote the [0001] and
[1000] direction in hexagonal crystal structure,
respectively................................................................................................ 5
Table 1.2. Fundamental properties of Cu2O. Unless indicated, all properties are at
room temperature....................................................................................... 7
Table 2.1. MOCVD process conditions for ZnO thin film......................................... 19
Table 2.2. PECVD process conditions for ZnO thin film........................................... 37
Table 3.1. Sputtering process conditions for ZnO thin film. The target is undoped
ZnO and substrates are soda lime glasses (microscope slides)................. 50
Table 3.2. Surface roughness and electrical resistivity of as-deposited and
thermally annealed films. For comparison, the values for a film
deposited at 350˚C are also presented. Film thickness is the sample
thicknesses which are used in resistivity measurement............................. 61
Table 5.1. The figures of merit for Cu2O/ZnO heterojunction solar cells. The
structure of solar cells is Au/Cu2O/(TiO2)/ZnO/Al-ZnO/ITO/Glass with
different ZnO thicknesses (65 nm, 130 nm, 200 nm). A fourth cell had a
10 nm thick TiO2 buffer layer between the Cu2O and ZnO layers............ 97
Table 5.2. Extracted saturation current activation energies for Cu2O/ZnO solar
cells. The structure of solar cells is Au/Cu2O/(TiO2)/ZnO/Al-ZnO
/ITO/Glass with different ZnO thicknesses, 65 nm, 130 nm, 200 nm and
TiO2 10 nm + ZnO 65 nm. The last sample has TiO2 buffer layer
between Cu2O and ZnO layers.................................................................. 112
xi
List of Figures
Figure 1.1. Hexagonal wurtzite crystal structures of ZnO........................................ 3
Figure 1.2. Cubic cuprite crystal structure of Cu2O................................................ 6
Figure 2.1. Hexagonal crystal structure unit cell and planes.................................... 13
Figure 2.2. Sequential process steps that lead to chemical vapor deposition
(CVD) of thin films on a substrate. The steps are (1) transport of the
precursor from gas inlet to reaction region, (2) chemical reaction in
gas phase to produce new reactive species, (3) transport of reactive
species to the substrate surface, (4) adsorption of reactive species on
the substrate and surface diffusion, (5) nucleation and growth to form
film and (6) desorption of by-products and pumping out from the
reaction region. This figure is adopted from reference 9....................... 15
Figure 2.3. Schematic of the CVD system for ZnO thin film deposition. Red lines
indicate heated lines............................................................................... 17
Figure 2.4. (a) CVD system for ZnO thin film deposition ; (b) shows the shower
head ; (c) shows the solid precursor tank where the bottom of the
precursor tank is surrounded by a heater................................................ 18
Figure 2.5. Arrhenius plot of the deposition rate versus the substrate temperature
(1000/T). The substrates were Si wafers. Solid line shows the linear
fit to the data between 573 K and 773 K................................................ 20
Figure 2.6. ZnO deposition rate versus Ar-to-O2 flow rate ratio. The substrate
temperature and process pressure were kept constant at 773 K and 5
Torr, respectively.................................................................................... 21
Figure 2.7. ZnO deposition rate versus the process pressure. The flow rates of the
carrier gas (Ar) and oxygen were kept at 200 sccm and 300 sccm,
respectively The substrate temperature was 773 K................................ 21
xii
Figure 2.8. X-ray diffraction from ZnO films deposited by MOCVD on various
substrates. (a) Glass, (b) (100) oriented Si, (c) (111) oriented Si, (d) a-
plane sapphire and (e) r-plane sapphire................................................. 23
Figure 2.9. XRD rocking curves of ZnO thin films deposited on Si (111), a-plane
sapphire and r-plane sapphire. The process pressure and substrate
temperature were kept constant at 5 Torr and 600 ˚C, respectively....... 24
Figure 2.10. SEM images of ZnO thin films deposited on (a) (111)-oriented Si,
(b) a-plane sapphire and (c) r-plane sapphire. The scale bars are 100
nm......................................................................................................... 25
Figure 2.11. XRD θ-2θ scan of ZnO film deposited on r-plane sapphire. The
process pressure and substrate temperature were 5 Torr and 600 ˚C,
respectively........................................................................................... 26
Figure 2.12. XRD in-plane φ-scan. (a) is the scan of (0006)Al2O3 peak from
sapphire substrate and (b) is the scan of ZnO(1010) peak form ZnO
film........................................................................................................ 27
Figure 2.13. Atomic arrangement of (a) sapphire (1102) (r-plane) and (b) ZnO
(1120) crystal planes, respectively. In-planar directions are denoted
on top of each diagram. Based on this epitaxial relationship, the
lattice mismatches between ZnO and sapphire are 1.53 % along ZnO
c-axis and 18.3 % perpendicular to the ZnO c-axis.............................. 28
Figure 2.14. Illustration of the in-plane φ-scan measurement. (a) An illustration of
the XRD geometry and the definition of the angle θ, φ and ψ. (b)
Shows the ZnO film on r-plane sapphire with (1120) plane parallel
to r-plane as determined in θ-2θ scan. (c) Shows the situation when
sample is tilted by 60° along with ψ direction..................................... 30
Figure 2.15. XRD rocking curves for ZnO(1120) peak of ZnO films deposited at
500˚C, 600°C and 700˚C...................................................................... 31
xiii
Figure 2.16. Temperature dependence of FWHM values of XRD rocking curves
(∆ω) and XRD diffraction peaks in the θ-2θ scan (∆2θ)...................... 31
Figure 2.17. SEM images of ZnO films deposited at (a) 500 ˚C, (b) 600 ˚C and
(c) 700˚C. SEM in (a) is the cross section image so that bottom is
sapphire substrate and upper is film. SEMs shown in (b) and (c) are
the top views. During deposition, Ar flow rate, O2 flow rate and
pressure are kept at 100 sccm, 300 sccm and 5.4 Torr, respectively.
The scale bars are 1 µm........................................................................ 32
Figure 2.18. AFM images of ZnO films deposited (a) at 400˚C, (b) at 500˚C and
(c) at 600˚C. The scan area is 1 µm × 1 µm......................................... 34
Figure 2.19. Dependence of the ZnO lattice parameter, a, on the deposition
temperature. Lattice parameters are calculated from the XRD
ZnO(1120) peak positions....................................................................... 34
Figure 2.20. SEM images of ZnO films deposited at step temperature profile. (a)
and (c) are the top view and cross section images of a film deposited
at 500˚C/2hrs - 600˚C/2hrs. (b) and (d) are the top view and cross
section images of a film deposited at 500˚C/2hrs - 700˚C/2hrs. The
scale bar is 1 µm.................................................................................. 36
Figure 2.21. Schematic of PECVD system for ZnO thin film deposition. Argon
and oxygen plasma is maintained during deposition by radio
frequency power.................................................................................. 38
Figure 2.22. SEM images of ZnO thin films deposited by PECVD at 600˚C and
3.5 Torr. The applied rf powers were (a) 20 W, (b) 30 W and (c) 40
W, respectively. The scale bars are 500 nm.......................................... 39
Figure 2.23. The cross sectional SEM images of ZnO films deposited by PECVD
with different thicknesses. The films were deposited at standard
conditions (Table 2.2) for (a) 30 min, (b) 60 min, (c) 90 min and (d)
120 min................................................................................................. 40
xiv
Figure 2.24. The plot of FWHM of XRD θ-2θ scans and rocking curves of ZnO
films versus film thickness. The measurements were conducted on
the same films in Figure 2.23................................................................ 41
Figure 2.25. (a) Carrier concentration and mobility and (b) resistivity as a
function of ZnO film thickness. Films were deposited by PECVD on
silicon under standard process conditions (Table 2.2). Film thickness
was varied by adjusting the deposition time : 30 min, 60 min and 120
min, respectively in the order of thickness........................................... 42
Figure 2.26. The SEM images of ZnO thin films deposited by PECVD on silicon.
The films were deposited under standard conditions (Table 2.2) but
at (a) 3.6 Torr and (b) 1.5 Torr.............................................................. 44
Figure 2.27. The XRD of films deposited under standard conditions (Table 2.2)
but at (a) 3.6 Torr and (b) 1.5 Torr. The measured films are same as
shown in Figure 2.26............................................................................. 44
Figure 2.28. The effect of pressure on the optical transmittance of thin ZnO films
deposited by PECVD on glass.............................................................. 45
Figure 3.1. A schematic of the magnetron sputtering system for ZnO thin film
deposition............................................................................................... 51
Figure 3.2. XRD from a ZnO thin film sputtered at room temperature on glass
substrate. The argon and oxygen flow rates were 19 and 1 sccm,
respectively............................................................................................. 52
Figure 3.3. SEM image of a ZnO film sputtered onto a glass substrate at room
temperature. The argon and oxygen flow rates were 19 and 1 sccm,
respectively and the deposition time was 7560 seconds........................ 52
Figure 3.4. The transmittance of ZnO films of varying thickness sputtered at
room temperature on glass. The argon and oxygen flow rates were 19
sccm and 1 sccm, respectively............................................................... 53
xv
Figure 3.5. The plot of lattice constant, c, and grain sizes of ZnO films versus the
concentration of O2 in feed gas. The grain sizes are calculated by
Scherrer’s formula from FWHM’s of (0002)ZnO peaks.......................... 55
Figure 3.6. The plot of film deposition rate versus the concentration of O2 in feed
gas........................................................................................................... 55
Figure 3.7. The plot of (a) lattice constant “c” and (b) grain sizes of as-deposited,
thermal annealing treated and oxygen plasma with thermal annealing
treated ZnO thin films versus the concentration of O2 in feed gas........ 57
Figure 3.8. (a) Lattice parameter, c and (b) grain size of as-deposited ZnO films
before and after 2 hour and 5 hour thermal annealing in presence of
an oxygen plasma as a function of film thickness................................. 58
Figure 3.9. The grain size and lattice parameter, c, of ZnO films as a function of
annealing temperature. The films were deposited on glass substrates
at room temperature while flowing argon and oxygen at 19 sccm and
1sccm, respectively. Annealing was conducted under vacuum for 1
hour........................................................................................................ 59
Figure 4.1. Schematic of the MOCVD system for Cu2O deposition........................ 66
Figure 4.2. (a) and (b) are typical XRD from Cu2O thin films deposited at 400˚C
on ZnO film. (a) shows XRD of a Cu2O film deposited on a smooth
ZnO surface. The RMS roughness was 3.5 nm and (b) shows XRD
from a Cu2O film deposited on a rough ZnO surface with RMS
roughness of 7.7 nm. (c) Typical SEM image of a Cu2O film
deposited by MOCVD at 400˚C. This film was deposited on ZnO
coated glass substrate at a total pressure of 2 Torr for 28 minutes........ 68
Figure 4.3. Arrhenius plot of the Cu2O deposition rate at a deposition pressure of
2 Torr. The linear line is the fitted line.................................................. 69
Figure 4.4. The plot of deposition rate versus process pressure. The films were
deposited at two substrate temperature, 350 and 400˚C........................ 69
xvi
Figure 4.5. (a) - (c) Digital photographs and (d) - (f) cross sectional SEM images
of Cu2O films deposited at substrate temperatures of 300 (a, d), 350
(b, e) and 400˚C (c, f), respectively. Scale bars are 500 nm. (g)
Optical absorbance of the films shown in (a) - (f)................................. 71
Figure 4.6. Tauc plot generated from the absorption data in Figure 4.5 for films
deposited at 350˚C and 400˚C................................................................ 72
Figure 4.7. XRD peaks from the films deposited at different substrate
temperatures 300, 350 and 400˚C. All the films were deposited under
a total pressure of 2 Torr and the film thicknesses were made similar
(222 nm, 208 nm and 243 nm respectively) by adjusting the
deposition times..................................................................................... 73
Figure 4.8. (a) Grain size and lattice parameter of the Cu2O films as a function of
deposition temperature. Equilibrium lattice constant of Cu2O is
4.267Å. (b) Carrier concentration, carrier mobility and (c) resistivity
as a function of deposition temperature................................................. 75
Figure 4.9. XRD from Cu2O films deposited at the same temperature (350˚C) as a
function of film thickness....................................................................... 76
Figure 4.10. (a) Grain size and lattice parameter of the Cu2O films as a function
of film thickness. (b) Carrier concentration, carrier mobility and (c)
resistivity as a function of film thickness. Deposition temperature
was 350˚C for all films......................................................................... 77
Figure 4.11. Atomic arrangements of (a) copper atoms on Cu2O (220) plane and
(b) oxygen atoms on ZnO (0002) plane............................................... 80
Figure 4.12. XRD from a Cu2O film deposited by MOCVD on ZnO which itself
was deposited on a-plane sapphire substrate by PEMOCVD. Cu2O
film was deposited at 400˚C and 2 Torr. ZnO film was deposited at
600˚C and 3.6 Torr............................................................................... 82
xvii
Figure 4.13. (a) Cubic Cu2O and (b) hexagonal ZnO unit cells illustrating that the
zone axis of the Cu2O (220) and (200) planes is the 001⟨ ⟩ direction
and that the zone axis of ZnO (0002) and (1011) planes is the
1210⟨ ⟩ direction. XRD φ-scan measurements (c) with 2θ fixed at
42.328˚ [Cu2O (200) diffraction] and (d) at 36.29˚
[ZnO (1011) diffraction]........................................................................ 83
Figure 5.1. Schematic of Cu2O/ZnO solar cell structure and a real sample picture.
For the measurement of J-V characteristics, positive bias was applied
to the Au pad in the forward bias case................................................... 89
Figure 5.2. Schematic of a solar cell equivalent circuit............................................ 91
Figure 5.3. Illustration of a virtual J-V characteristic for a solar cell under
illumination. Vm and Jm are the maximum voltage and maximum
current density, respectively. Maximum power (Pmax) is produced in
the ranage 0 < V < Voc and 0 < J < Jsc. Fill factor is the ratio of |Jm ×
Vm| to |Jsc × Voc|.................................................................................... 92
Figure 5.4. Equipment for measuring the temperature dependent J-V
characteristics of solar cells. (a) is the photograph of the overall
system. (b) is the schematic of the measurement set up........................ 95
Figure 5.5. J-V characteristics of two Cu2O/ZnO heterojunction solar cells. The
cell structures are Au/Cu2O/ZnO/Al-ZnO/ITO/Glass with 65 nm thick
ZnO and 130 nm thick ZnO. The “Dark” and “Light” J-V
characteristics were measured in the dark and the under illumination,
respectively............................................................................................ 97
Figure 5.6. The spectrum of (a) LHE, (b) IQE and (c) IPCE of Cu2O/ZnO solar
cell with 130 nm thick ZnO................................................................... 99
Figure 5.7. Dark J-V characteristics of Cu2O/ZnO solar cells as a function of
temperature between 160 K and 295 K. The thin film stack that
comprised this solar cell was Au/Cu2O/ZnO/AZO/ITO/Glass: (a) 65
nm thick ZnO (b) 130 nm thick ZnO..................................................... 104
xviii
Figure 5.8. (a) Dark J-V characteristics (log - log scale) of a Cu2O/ZnO solar cell
with 130 nm thick ZnO as a function of temperature. The solid lines
are power law fits to the data: fits for only one set of data is shown in
(a). The behavior is ohmic at low bias values with J ~ V. The
exponent, m, in the power law fits depends on the temperature at high
bias values as shown in (b). (c) The “crossover” plot for the data in
(a): all power law fits are shown as lines and intersect at the crossover
voltage. (d) J-V characteristics (log - log scale) of the same solar cell
in (a) but under illumination. The solid line is the power law fit to one
set of data at 295 K. Fits for other sets are not shown for clarity but
the exponent, m, in the power law fits is nearly independent of
temperature in the high bias region....................................................... 107
Figure 5.9. Semi-log plot of short circuit current (Jsc) versus open circuit voltage
(Voc) as a function of temperature and illumination intensity for a
Cu2O/ZnO heterojunction solar cell with 130 nm thick ZnO layer.
Solid lines are fits to the data using equation (5.9) to extract Al and Jol. 109
Figure 5.10. (a) Open circuit voltage (Voc) as a function of temperature for solar
cells listed in Table 5.2. The lines are linear fits to the data and show
extrapolation of Voc to T = 0 K to determine the activation energy in
equation (5.8) from the T = 0 K intercept. (b) The plot of the AllnJol
versus reciprocal temperature for the same solar cells as in (a). The
lines are fits to the data using equation (5.10). The activation energy
for each cell is determined from the slopes of these lines................... 111
Figure 5.11. Band diagram of the Cu2O/ZnO heterojunction. EC, EV, EF, and Ea
are the conduction-band-edge, valence-band-edge, Fermi-level, and
activation energies, respectively. Most of the band bending is shown
to be on the Cu2O side because of lower doping in Cu2O compared
to the ZnO side. Paths 1 and 2 indicate two possible hole
recombination pathways through interface defects. Activation energy 115
xix
(Ea) is the barrier that the holes have to overcome to recombine. The
approximate energy levels are based on the properties of bulk Cu2O
and ZnO. Electron affinities: 3.2 eV (Cu2O) and 4.2 eV (ZnO); band
gaps: 2.17 eV (Cu2O ) and 3.44 eV (ZnO); dielectric constants: 7.11
(Cu2O) and 7.8 (ZnO), electron effective mass: 0.275mo (ZnO) and
hole effective mass in 0.58m0 (Cu2O) - where mo is free electron
mass. Measured carrier concentrations of ~1015 cm-3 (Cu2O) and ~ 2
× 1017 cm-3 (ZnO) were used...............................................................
Figure 5.12. (a) Diode ideality factor versus reciprocal temperature for J-V
characteristics acquired in the dark from solar cells made using 65
nm thick ZnO with and without a 10 nm thick TiO2 buffer layer. The
solid line for the solar cell without the buffer layer is a fit to equation
(5.16). The solid line for the solar cell with the buffer layer is drawn
to guide the eye. (b) Diode ideality factor versus reciprocal
temperature for J-V characteristics acquired under 100 mW/cm2
illumination from solar cells listed in Table 5.2................................... 118
Figure 5.13. Copper, zinc and indium concentration profile measured using
Auger depth profiling........................................................................... 120
Figure 6.1. Band gap diagram of Cu2O, ZnS and ZnO. Ec, Ef and Ev are the
conduction band edge, fermi-level and valence band edge,
respectively. The dopant binding energy is arbitrarily selected
therefore the positions of EF are arbitrary. The electron affinity and
band gap of each materials are come from the references 6,7,8,9......... 127
Figure 6.2. Cross sectional SEM image of ZnO nanowires grown on ITO coated
glass substrate with and without Cu2O film deposited by MOCVD.
(a) is the ZnO nanowires grown on ITO glass before the Cu2O
deposition. (b) is the after the Cu2O deposition..................................... 129
1
Chapter 1
Introduction
1.1 Cu2O/ZnO heterojunction and solar cell application
Metal oxides exhibit a wide range of functional properties depending on their
crystal structure and bonding between the metal cation and oxygen.1,2 Indeed, the
electrical properties of metal oxides range from insulating to highly conducting like a
metal, or even superconducting. Some metal oxides also exhibit magnetic properties such
as ferromagnetism or ferrimagnetism. Because of this diverse functionality, metal oxides
have become one of the most fascinating inorganic materials in device applications such
as light emitting diodes, field effect transistors, solar cells and spintronic devices.3,4,5,6,7
Among the metal oxides, ZnO and Cu2O are attractive because the metals are abundant
on earth, inexpensive and non-toxic. Moreover, these oxides have useful optical and
electrical properties suitable for a wide variety of electrical devices.
As synthesized ZnO is almost always an n-type semiconductor with high mobility
(10 ~ 200 cm2 / V sec) in thin high quality films at room temperature. ZnO is a wide band
gap (3.3 eV)8 semiconductor and therefore, it has been considered as a potential material
2
for transparent electrode and electronics9,10 and widely used in solar cells as window
layer and transparent thin film transistors.11 As synthesized Cu2O is a p-type
semiconductor with band gap of 2.17 eV.12 Although the band gap is larger than the
theoretical optimum band gap (~ 1.5 eV)13 for solar cells, Cu2O has several advantages as
a light absorbing layer in solar cells such as large absorption coefficient in the visible
region (~ 104 cm-1) and a long minority carrier diffusion length (1 ~ 4 µm).14 Based on
the fact that high quality Cu2O can be fabricated by simple thermal oxidation at around
1000˚C from pure copper, there have been extensive studies to use Cu2O in low cost
photovoltaic power generation15,16,17,18 for the several decades by simply forming a Cu-
Cu2O schottky junction but the overall efficiency has not reached above 2 % so far.
Recently, thin film solar cells based on the heterojunction between p-type Cu2O and n-
type ZnO has received attention because of the successful demonstration of ZnO and
Cu2O thin film deposition by various methods such as molecular beam epitaxy (MBE),19
sputtering,20,21 chemical vapor deposition (CVD)22 and electrochemical deposition.23 The
performance of thin film heterojunction solar cells depends on the electrical and optical
properties of each layer which comprises the device and the quality of the interface
between the films that form the junction. These critical properties in turn depend on the
thin film deposition method and the deposition process conditions. Therefore, an in depth
investigation of the relationship between the film properties and carrier transport
mechanisms through the junction interface and the deposition process conditions is
important and needed.
3
1.2 Review of relevant ZnO and Cu2O properties
1.2.1 ZnO properties
ZnO is a II-VI semiconductor and crystallizes in three phases, wurtzite
(hexagonal), zinc-blende (cubic) and rocksalt (cubic).8 The stable phase under ambient
conditions is wurtzite where each Zn cation atom is bonded with four O anion atoms
which form the corners of a tetrahedron. Zinc-blende can be formed on cubic structured
substrates and rocksalt can be made at high pressure (~ 100 kbar).24 A schematic diagram
of ZnO wurtzite crystal structure is shown in Figure 1.1 and selected fundamental
properties of ZnO are listed in Table 1.1.8,25
Figure 1.1. Hexagonal wurtzite crystal structures of ZnO.
4
ZnO is transparent in the visible region of electromagnetic spectrum and the
absorption increases abruptly in the ultraviolet region for photon energies greater than 3.3
eV, the direct band gap energy. ZnO thin films are usually n-type even in the absence of
intentional doping. It was reported that the main source of intrinsic donors are point
defects such as oxygen vacancies or zinc interstitials26 but recent calculations by Van de
Walle shows that these kinds of defects form deep level donors which are hard to ionize
and unlikely to donate electrons to the conduction band at room temperature. Van de
Walle argued that the electrons originated from hydrogen impurities in ZnO and showed
that H in ZnO is a shallow donor.27 This debate is still unsettled but experimental fact is
that exposing ZnO to hydrogen increases electron concentration whereas exposure to
oxygen decreases it.28
Native donors such as oxygen vacancies, zinc interstitials and hydrogen in ZnO
are unstable and their concentrations change with time and temperature. Concentrations
of these native donors also depend on film deposition methods and processing conditions.
To improve electrical properties of films and make them stable, intentional doping is
used. The III or V elements in the periodic table can be used as n-type and p-type dopants
for ZnO, respectively. For n-type doping B, Al, Ga and In etc. are used and for p-type
doping N and P etc. are used. However, ZnO films almost always show n-type
characteristic regardless of the type of dopants used.29 The reason for this is that p-type
dopant such as N and P is easily compensated by native defects.30 So the difficulty in
making a high quality p-type ZnO film remains a major obstacle for the widespread use
of ZnO in opto-electronic device applications.
5
Table 1.1. Fundamental properties of ZnO. Unless indicated otherwise, all properties are
at room temperature. “c” and “a” denote the [0001] and [1000] direction in hexagonal
crystal structure, respectively.8,25
Properties Values
Density (g/cm3) 5.67
Melting point (K) 2242
Crystal structure Wurtzite (Hexagonal) (Space group : P63mc - C4
6v)
Lattice constant (Å) a = 3.2495; c = 5.2069
Band gap (eV) 3.3 (Direct)
Electron Hall mobility (cm2/V sec) n cμ ⊥ = 70 (film); n cμ ⊥ = 150 (bulk);
n cμ = 170 (bulk)
Dielectric Constant (0) cε ⊥ = 7.8; ( ) cε ∞ ⊥ = 3.7;
(0) cε = 8.75; ( ) cε ∞ = 3.75
1.2.2 Cu2O properties
Copper oxide has two stable phases, cuprous oxide (Cu2O) and cupric oxide
(CuO). The Cu2O phase has reddish color with direct band gap of 2.17 eV and CuO has
dark gray color with an indirect band gap of 1.4 eV. The property difference between
Cu2O and CuO is mainly due to the different Cu valence states. Copper is in Cu+ state in
6
Cu2O and in Cu2+ state in CuO.31,32 Cu2O shows spontaneous p-type semiconductor
characteristics and it was shown that this p-type conductivity comes from cation-
deficiency which is mainly due to the copper vacancies rather than oxygen interstitials or
other defects.33 Doping mechanism of Cu2O is not well understood so far but several
elements are known as efficient p-type dopants such as chlorine (Cl), silicon (Si) and
nitrogen (N).34,35,36 N-type doping has not been successfully demonstrated yet.
The crystal structure of Cu2O is the cubic cuprite structure with lattice parameter
4.2696 Å. A unit cell has four copper atoms and two oxygen atoms and the oxygen atoms
form body centered cubic (bcc) lattices and the copper atoms form face centered cubic
(fcc) lattices as shown in Figure 1.2. Some fundamental properties of Cu2O are listed in
Table 1.2.
Figure 1.2. Cubic cuprite crystal structure of Cu2O.
7
Table 1.2. Fundamental properties of Cu2O. Unless indicated, all properties are at room
temperature.12,37
Properties Values
Density (g/cm3) 5.749 ~ 6.14 (variation due to the presence of voids in most synthetic materials)
Melting temperature (K) 1508
Crystal structure Cuprite (Cubic) (Space group : Pn3m - Oh
4)
Lattice constant (Å) a = 4.2696
Band gap (eV) 2.17 (direct)
Hole Hall mobility (cm2/V sec) µ = 70
Dielectric constant ε(0) = 7.11; ε(∞) = 6.46
1.3 Thesis organization
Following introduction of chapter 1, thin film deposition of ZnO by metal organic
chemical vapor deposition (MOCVD) and plasma enhanced chemical vapor deposition
(PECVD) are described in chapter 2. Specifically, in chapter 2, the dependence of ZnO
thin film morphology on the substrate materials and orientation was investigated. I found
that ZnO thin film shows epitaxy on r-plane sapphire with 2 3
(1120) (1102)ZnO Al O and
8
2 3[0001] [1120]ZnO Al O⊥ and this is also discussed in detail in chapter 2.
In chapter 3, I discuss the properties of ZnO thin films deposited by radio
frequency magnetron sputtering on glass substrates. Specifically the electrical, structural
and optical properties of sputtered ZnO on glass are described. The effect of key process
factors such as oxygen flow rate and substrate temperature was investigated. Especially I
show how the various post-annealing treatments such as thermal annealing and oxygen
plasma exposure affect the film properties.
In chapter 4, I describe deposition of thin Cu2O films by MOCVD. Cu2O was
deposited on ZnO thin films to form heterojunctions and the dependence of film
properties on substrate temperature and film thickness was investigated. Although Cu2O
and ZnO have different crystal structures (cubic and hexagonal), Cu2O grows epitaxially
on ZnO and I discuss the reasons for this in detail in this chapter.
In chapter 5, I describe the fabrication of Cu2O/ZnO heterojunction thin film solar
cells and show the results of photovoltaic characterization. In order to investigate the
carrier transport mechanism in detail, current-voltage (J-V) characteristics in the dark and
under illumination were collected and analyzed as a function of temperature between 100
K - 300 K. These measurements are presented and discussed in detail in this chapter.
Finally, Chapter 6 includes summary and main conclusions along with future directions.
9
References
1 David P. Norton, Mater. Sci. Eng. R 43, 139 (2004).
2 C. N. R. Rao, B. Raveau, Transition Metal Oxides, Wiley-VCH, New York, 1998.
3 D. C. Look, Mat. Sci. Eng. B80, 383 (2001).
4 J. F. Wager, Science 300, 1245 (2003).
5 K. Koike, I. Nakashima, K. Hashimoto, S. Sasa, M. Inoue, M. Yano, Appl. Phys. Lett.
87, 112106 (2005).
6 K. Akimoto, S. Ishizuka, M. Yanagita, Y. Nawa, G. K. Paul, T. Sakurai, Sol. Energy 80,
715 (2006).
7 S. J. Pearton, W. H. Heo, M. Ivill, D. P. Norton, T. Steiner, Semicond. Sci. Tech. 19,
R59 (2004).
8 U. Ozgur, Ya .I. Alivov, C. Liu, A,Teke, M.A. Reshchikov, S. Dogan, V. Avrutin, S.J.
Cho, H. Morkoc, J. Appl. Phys. 98, 041301 (2005).
9 T. Minami, MRS Bull. 38 , August (2000).
10 R. G. Gordon, MRS Bull. 52, August (2000).
11 R. L. Hoffman, B. J. Norris, J. F. Wager, Appl. Phys. Lett. 82, 733 (2003).
12 Otfried Madelung, Semiconductors : Data Handbook, Springer, 3rd edition, 2004.
13 J. J. Loferski, J. Appl. Phys. 27, 777 (1956).
14 C. A. Dimitriadis, L. Papadimitriou, N. A. Economou, J. Mat. Sci. Lett. 2, 691 (1983).
15 L. C. Olsen, F. W. Addis, W. Miller, Sol. Cells 7, 247 (1982-1983).
16 J. Herion, E. A. Niekisch, G. Scharl, Sol. Energ. Mater. 4, 101 (1980).
17 L. Papadimitriou, N. A. Economou D. Trivich, Sol. Cells 3, 73 (1981).
10
18 A. E. Rakhshani, Solid State Electron. 29, 7 (1986).
19 J. Q. Xie, J. W. Dong, A. Osinsky, P. P. Chow, Y. W. Heo, D. P. Norton, S. J. Pearton,
X. Y. Dong, C. Adelmann, C. J. Palmstrom, Mater. Res. Soc. Symp. P. 891, 407
(2006).
20 K. Ellmer, J. Phys. D. Appl. Phys. 34, 3097 (2001).
21 S. Ishizuka, K. Suzuki, Y. Okamoto, M. Yanagita, T. Sakurai, K. Akimoto, N.
Fujiwara, H. Kobayashi, K. Matsubara, S. Niki, Phys. Status. Solidi. 1, 1067 (2004).
22 R. G. Gordon, Mater. Res. Soc. Symp. P. 426, 419 (1996).
23 M. Izaki, T. Shinagawa, K. Mizuno, Y. Ida, M. Inaba, A. Tasaka, J. Phys. D Appl.
Phys. 40, 3326 (2007).
24 C. H. Bates, W. B. White, R. Roy, Science 137, 993 (1962).
25 Satischandra B. Ogale, Thin films and Heterostructures for oxide electronics, Chapter
10, 303 (Springer, 2005).
26 D. C. Look, J. W. Hemsky, J. R. Sizelove, Phys. Rev. Lett. 82, 2552 (1999).
27 Chris G. Van de Walle, Phy. Rev. Lett. 85, 1012 (2000).
28 C. A. Wolden, T. M. Bames, J. B. Baxter, E. S. Aydil, J. Appl. Phys. 97, 043522
(2005).
29 B. Claflin, D. C. Look, S. J. Park, G. Cantwell, J. Cryst. Growth 287, 16 (2006).
30 Y. W. Heo, S. J. Park, K. Ip, S. J. Pearton, D. P. Norton, Appl. Phys. Lett. 83, 1128
(2003).
31 W. Y. Ching, Y. N. Xu, K. W. Wong, Phys. Rev. B 40, 7684 (1989).
32 T. ITO, H. Yamaguchi, K. Okabe, T. Masumi, J. Mater. Sci. 33, 3555 (1998).
11
33 H. Raebiger, S. Lany, A. Zunger, Phys. Rev. B 76, 045209 (2007).
34 A. O. Musa, T. Akomolafe, M. J. Carter, Sol. Energ. Mat. Sol. C. 51, 305 (1998).
35 S. Ishizuka, S. Kato, Y. Okamoto, K. Akimoto, Appl. Phys. Lett. 80, 950 (2002).
36 Y. Nakano, S. Saeki, T. Morikawa, Appl. Phys. Lett. 94, 022111 (2009).
37 G. P. Pollack, D. Trivich, J. Appl. Phys. 46, 163 (1975).
12
Chapter 2
Properties of ZnO thin films deposited by metal organic chemical vapor deposition
2.1 Introduction
In this chapter, properties of ZnO thin films deposited by metal organic chemical
vapor deposition (MOCVD) are described. ZnO thin films can be deposited by various
techniques such as radio-frequency magnetron sputtering,1 molecular beam epitaxy
(MBE),2 pulsed laser deposition (PLD),3 chemical vapor deposition (CVD),4
electrochemical deposition5 and spray pyrolysis.6 Applications in opto-electronic devices
require high quality epitaxial thin films and among these methods, MBE, PLD and CVD
can produce epitaxial ZnO films. CVD is particularly attractive because it can be scaled
up for mass production. For this reason, it is one of the most popular deposition
techniques in the semiconductor industry.
ZnO has wurtzite hexagonal crystal structure and tends to grow in columnar
morphology because growth rate in c-axis direction is much faster than the growth rate of
basal planes.7 Because of this anisotropic growth and columnar morphology, it is difficult
to deposit films with smooth surface. However, many device applications require thin
films with smooth surface so that abrupt and smooth heterojunctions can be formed. I
13
deposited ZnO thin films on various substrates including amorphous glass, silicon and
sapphire (a-plane and r-plane, Figure 2.1) by MOCVD and investigated the dependence
of the structure and morphology of ZnO thin films on the substrate. All the zinc oxide
thin films deposited on amorphous glass, silicon and a-plane sapphire shows textured
structure with (0002), (c-axis) parallel to the substrate normal but the films deposited on
r-plane sapphire have the c-axis parallel to substrate surface. ZnO has epitaxial
relationship with r-plane sapphire and this is described in detail in this chapter including
the dependence of morphology on the deposition temperature.
Figure 2.1. Hexagonal crystal structure unit cell and planes.
To control the morphology of ZnO thin films, plasma enhanced chemical vapor
deposition (PECVD) was also investigated. PECVD has several advantages that may
improve the deposited film quality and morphology.8 Specifically, high energetic particle
14
bombardment from plasma can affect surface reactions and dissociate metal organic
precursors. The dependence of the structure and morphology of ZnO films deposited by
PECVD on process parameters such as applied power and process pressure was
investigated and this is described in this chapter.
2.2 Metal organic chemical vapor deposition of ZnO thin films
2.2.1 Chemical vapor deposition (CVD) process
Chemical vapor deposition (CVD) makes use of chemical reactions on a hot
surface. The steps that lead to the deposition of a thin film in CVD are shown in figure
2.2.9 Typically the film precursors are delivered to the deposition chamber by a carrier
gas and react with other gases on the heated substrate surface to nucleate and grow the
desired film. The volatile by-products of the reactions are pumped out.
Several important factors affect the quality of the film deposited by CVD. These
include the properties of the precursor, the substrate, the deposition temperature, the
process pressure, the carrier gas flow rate and the chamber geometry. To deposit ZnO
films by CVD, a zinc containing precursor and an oxygen source are needed. Metal-
organic compounds such as diethyl zinc [Zn(C2H5)2, DEZ], dimethyl zinc [Zn(CH3)2,
DMZ], zinc acetate [Zn(CH3COO)2] and zinc acetylacetonate [Zn(C5H7O2)2, Zn(acac)2]
have been used as zinc precursors. Oxygen sources that have been used in conjunction
with these zinc precursors include O2, CO2, H2O, N2O and NO2.10
15
Figure 2.2. Sequential process steps that lead to chemical vapor deposition (CVD) of thin
films on a substrate. The steps are (1) transport of the precursor from gas inlet to reaction
region, (2) chemical reaction in gas phase to produce new reactive species, (3) transport
of reactive species to the substrate surface, (4) adsorption of reactive species on the
substrate and surface diffusion, (5) nucleation and growth to form film and (6) desorption
of by-products and pumping out from the reaction region. This figure is adopted from
reference 9.
Among the zinc precursors, DEZ is used most widely because it leads to high
quality films with O2 as the oxidizing gas and DEZ vapor flow and hence the deposition
rate can be controlled easily. However, DEZ is pyrophoric which makes its use and
delivery difficult. Instead, I used a solid precursor, zinc acetylacetonate (Zn(C5H7O2)2,
Zn(acac)2) because it is stable in ambient conditions and does not react with atmospheric
oxygen or water. Thus, the manipulation of Zn(acac)2 is easier than DEZ.
16
2.2.2 Chemical vapor deposition (CVD) system
I designed and set up a high vacuum CVD system for ZnO thin film deposition. A
schematic of this system is shown in Figure 2.3 and photographs are shown in Figure 2.4.
The 4-inch diameter cylindrical deposition chamber is evacuated through a side port
located just below the sample stage by a turbo molecular pump (TMP) which provides a
base pressure of 3 × 10-7 Torr. Rotary mechanical pumps are used for TMP backing and
initial chamber pumping from atmosphere to around 50 mTorr. The flow rate of gases
(argon for carrier gas, O2 for process and N2 for venting) are controlled by mass flow
controller (MFC) from zero to 3000 standard cubic centimeter per minute (sccm).
Substrate is placed on the sample stage which can be heated up to 1000˚C by resistance
heater and the substrate temperature is controlled by K-type thermocouple and close loop
feedback controller.
The precursor [Zn(acac)2] is placed in a stainless steel tank [Figure 2.4 (c)] with
one inlet for the carrier gas (Ar) and an outlet that led to the deposition chamber. The
precursor tank is surrounded by a band heater and a thermocouple is attached at the
bottom of the tank for independent temperature control. Sublimed precursors flow
downward from the top flange through a showerhead. During the deposition, the valve in
front of TMP is closed and pumping through the rotary pump is used. Process pressure is
adjusted using a butterfly valve located between the chamber and the rotary pump in the
range of 1 - 10 Torr and a capacitance manometer gauge is used for the accurate
measurement of the process pressure. The precursor and oxygen gases are fed into the
deposition chamber through separate lines to prevent the premature reaction between the
two gases. The gas line that connects the precursor tank to the showerhead is heated to
17
150˚C to prevent the condensation of sublimed precursor.
Figure 2.3. Schematic of the CVD system for ZnO thin film deposition. Red lines
indicate heated lines.
18
Figure 2.4. (a) CVD system for ZnO thin film deposition ; (b) shows the shower head ;
(c) shows the solid precursor tank where the bottom of the precursor tank is surrounded
by a heater.
19
2.2.3 ZnO film deposition kinetics
The deposition rate of ZnO thin films by MOCVD from Zn(acac)2 and oxygen on
silicon substrate was investigated. The process conditions used in this investigation are
summarized in Table 2.1. Figure 2.5 shows dependence of the deposition rate on the
substrate temperature between 573 K and 873 K. The deposition rate below 773 K shows
Arrhenius behavior indicating that the process may be surface-reaction-limited. The slope
of the deposition rate versus 1/T line is reduced above 773 K which indicates the process
may become mass-transport-limited at high temperature. From the slope of the line in the
low temperature region (573 K ~ 773 K), an activation energy of 96 kJ/mol is extracted.
Table 2.1. MOCVD process conditions for ZnO thin film.
Process condition Values
Deposition temperature (K) 573 - 873
Precursor sublime temperature (˚C) 90 - 110
Carrier gas (Ar) flow rate (sccm) 10 - 200
O2 flow rate (sccm) 300
Process pressure (Torr) 1 - 10
20
1.0 1.2 1.4 1.6 1.8
0.1
1
10
Dep
ositi
on ra
te (Å
/min
)
1000/T (1/K)
900 800 700 600 Temperature (K)
Figure 2.5. Arrhenius plot of the deposition rate versus the substrate temperature
(1000/T). The substrates were Si wafers. Solid line shows the linear fit to the data
between 573 K and 773 K.
Figure 2.6 shows the dependence of the deposition rate on the ratio of the carrier
gas (Ar) flow rate to the oxygen gas flow rate at constant process pressure (~ 5 Torr) and
constant substrate temperature (773 K). The Ar to O2 flow rate ratio is a measure of the
ratio of zinc precursor to O2 partial pressure in the reactor. Increasing this ratio increases
both the Zn precursor flow rate into the chamber and the ratio of the Zn precursor partial
pressure to O2 partial pressure. The deposition rate increases with Ar to O2 flow rate ratio
up to ~ 0.35 and then saturates. This behavior indicates that the process may be limited
by zinc precursor partial pressure when the Ar-to-O2 flow rate ratio is below 0.35. The
dependence of deposition rate on total process pressure (Figure 2.7) shows that it has a
maximum value at around 5 Torr.
21
0.0 0.2 0.4 0.6 0.80
2
4
6
8
10
Dep
ositi
on ra
te (Å
/min
)
Ar/O2 flow rate ratio
Figure 2.6. ZnO deposition rate versus Ar-to-O2 flow rate ratio. The substrate
temperature and process pressure were kept constant at 773 K and 5 Torr, respectively.
0 2 4 6 8 10 120
2
4
6
8
Dep
ositi
on ra
te (Å
/min
)
Process pressure (Torr)
Figure 2.7. ZnO deposition rate versus the process pressure. The flow rates of the carrier
gas (Ar) and oxygen were kept at 200 sccm and 300 sccm, respectively. The substrate
temperature was 773 K.
22
2.3 The dependence of ZnO film morphology on the substrates
The X-ray diffractions of ZnO thin films deposited by MOCVD on various
substrates are shown in Figure 2.8. The substrates included amorphous glass, (100)
oriented Si, (111) oriented Si, a-plane sapphire and r-plane sapphire. With the exception
of r-plane sapphire, the XRD from films deposited on all these substrates show only the
substrate peaks and the (0002)ZnO peak. The film deposited on r-plane sapphire substrate
shows (1120) ZnO diffraction in addition to the substrate peaks. Ohotomo et al. also
observed the same in ZnO thin films grown by pulsed laser deposition (PLD). Thin ZnO
films grown by PLD showed that [0002] axis is parallel to the substrate surface normal
except on r-plane sapphire.3
The fact that ZnO has a preferred growth direction regardless of substrates
indicates that this is not due to the epitaxy between the film and the substrates but is an
intrinsic tendency of ZnO to grow anisotropically with its c-axis oriented parallel to the
surface normal. ZnO has wurtzite hexagonal crystal structure and the surface energies are
anisotropic. Specifically, the surface formation energies are higher in the order of (0002),
(1011) and (1010) planes. The plane with lower formation energy is likely to grow larger
in size, while the planes with higher formation energies tend to minimize their area
during film growth. Therefore, the fastest growth direction will be the normal direction of
the plane with the highest formation energy. This makes the ZnO [0002] direction (c-
axis) the fastest growth direction.11
23
2x102
1x105
4x102
2x104
20 30 40 50 60 700
2x104
(0002)ZnO
(004)Si(0002)ZnO
Inte
nsity
(Cou
nts)
(111)Si (0002)ZnO
(1120)Al2O3
(0002)ZnO
2θ (degrees)
(1102)Al2O3
(2204)Al2O3
(1120)ZnO
Figure 2.8. X-ray diffraction from ZnO films deposited by MOCVD on various
substrates. (a) Glass, (b) (100) oriented Si, (c) (111) oriented Si, (d) a-plane sapphire and
(e) r-plane sapphire.
ZnO and sapphire have the same crystal structure (hexagonal wurtzite) while Si
has a different (cubic) structure. Thus, we would expect that the ZnO films deposited on
sapphire to have better crystallinity than those deposited on Si. Indeed, this expectation is
24
consistent with the XRD of film deposited on sapphire and Si. Figure 2.9 shows the XRD
rocking curves for thin ZnO films deposited on (111) oriented Si, a-plane sapphire and r-
plane sapphire. The full widths of the rocking curves at half the maximum intensity
(FWHM) are 6.7˚, 1˚ and 0.5˚ for films grown on (111)-oriented Si, a-plane sapphire and
r-plane sapphire, respectively. A smaller FWHM value of XRD rocking curves indicates
better crystallinity. Hence, the highest crystallinity films are grown on r-plane sapphire
followed by a-plane sapphire and (111)-oriented silicon. These results are also consistent
with the morphologies seen in SEM images displayed in Figure 2.10. These SEM images
show that the ZnO films on silicon substrates have a rough surface morphology while the
film deposited on a-plane and r-plane sapphire substrates show columnar growth and
smooth and dense morphology, respectively.
-2 -1 0 1 20.0
0.2
0.4
0.6
0.8
1.0
Si (111)
Rel
ativ
e in
tens
ity (a
.u.)
ω (deg)
a - Sapphirer - Sapphire
Figure 2.9. XRD rocking curves of ZnO thin films deposited on Si (111), a-plane
sapphire and r-plane sapphire. The process pressure and substrate temperature were kept
constant at 5 Torr and 600 ˚C, respectively.
25
Figure 2.10. SEM images of ZnO thin films deposited on (a) (111)-oriented Si, (b) a-
plane sapphire and (c) r-plane sapphire. The scale bars are 100 nm.
2.4 Structural properties of ZnO thin films grown on r-plane
sapphire
Since ZnO films tend to grow in columnar structures on many substrates with c-
axis perpendicular to the substrate surface, it is difficult to deposit films with smooth
surface.7 In contrast, ZnO films grow on r-plane sapphire with their c-axis, the preferred
growth direction, parallel to the substrate surface. Resulting films exhibit smooth
surfaces. In this section, structural properties of ZnO films on r-plane sapphire including
epitaxial relationship and surface morphology are described. Specifically I focus on the
substrate temperature between 400˚C and 700˚C.
The epitaxial relationship between films and substrates was characterized using
XRD θ-2θ scan and φ-scan measurements and crystal quality was examined through
XRD rocking curve measurements. Surface roughness was measured by atomic force
microscopy (AFM) and surface morphology was investigated by SEM.
(a) (b) (c)
26
2.4.1 Epitaxy between ZnO and r-plane sapphire
Figure 2.11 shows the result of XRD from a film deposited on r-plane sapphire at
600 ˚C. All other films deposited on r-plane sapphire between 400 ˚C ~ 700 ˚C showed
same XRD pattern. The positions of the three peaks are at 25.6˚, 52.6˚ and 56.7˚ which
correspond to the 2 3Al O(1102) ,
2 3Al O(2204) and ZnO(1120) , respectively. This indicates that
one component of the epitaxial orientation relationship is 2 3ZnO Al O(1120) (1102) .
20 30 40 50 60 700.0
0.4
0.8
1.2
1.6
2.0
Inte
nsity
(104 c
ount
s)
2θ (deg)
Figure 2.11. XRD θ-2θ scan of ZnO film deposited on r-plane sapphire. The process
pressure and substrate temperature were 5 Torr and 600 ˚C, respectively.
In-plane orientation of the ZnO films on r-plane sapphire was investigated using
XRD φ-scan measurements. Two φ-scans were conducted, one scan with 2θ fixed at 41.7˚
[(0006)Al2O3 diffraction] and another with 2θ fixed at 31.8˚ [ ZnO(1010) diffraction]. These
two φ-scans are shown in Figure 2.12. The (0006)Al2O3 diffraction peak from sapphire
2 3Al O(1102)2 3Al O(2204)
ZnO(1120)
27
substrate is separated from the two ZnO(1010) diffraction peaks by 90˚. The zone axis of
sapphire (0006) and (1102) planes is the [1120] direction, whereas the zone axis of ZnO
(1010) and (1120) planes is the [0001] direction which leads to the conclusion that
sapphire [1120] must be perpendicular to the ZnO [0001]. Therefore, combining this
result with the result of θ-2θ scan, 2 3ZnO Al O(1120) (1102) the complete epitaxial relation
between r-plane sapphire and ZnO is determined to be 2 3ZnO Al O(1120) (1102) and
2 3ZnO Al O[0001] [1120]⊥ .
-180 -90 0 90 180
0
1
2
φ (deg)
Inte
nsity
(105 c
ount
s)
Figure 2.12. XRD in-plane φ-scan. (a) is the scan of (0006)Al2O3 peak from sapphire substrate and (b) is the scan of ZnO(1010) peak form ZnO film.
This epitaxial relationship agrees with the growth mechanism proposed by Kung
et al.12 who developed a model of the epitaxial growth of wurtzite type thin films on r-
2 3Al O(0006)
ZnO(1010)
(a)
(b)
28
plane sapphire. According to this model, the surface of the r-plane sapphire is oxygen-
terminated and the oxygen atoms are aligned on the surface along the [1101] direction as
shown in Figure 2.13. The oxygen atoms on ZnO(1120) plane are also aligned along the
[0001] direction and when ZnO films are deposited on r-plane sapphire, rows of oxygen
atoms are deposited along the valley between oxygen ridges on sapphire substrate.
Therefore, ZnO films grow with c-axis parallel to the sapphire [1101] which is equivalent
to 2 3ZnO Al O[0001] [1120]⊥ .
Figure 2.13. Atomic arrangement of (a) sapphire (1102) (r-plane) and (b) ZnO (1120)
crystal planes, respectively. In-planar directions are denoted on top of each diagram.
Based on this epitaxial relationship, the lattice mismatches between ZnO and sapphire are
1.53 % along ZnO c-axis and 18.3 % perpendicular to the ZnO c-axis.
Figure 2.14 illustrates the XRD in-plane scan for ZnO(1010) plane used in this
experiment. Figure 2.14 (a) and (b) shows the XRD geometry and the geometry of the
(a) (b)
29
ZnO film with the ZnO(1120) plane parallel to the r-plane sapphire. Initially, the film
surface is parallel to the X-ray incident plane which is defined by the incident and
diffracted X-ray beams. Starting with this configuration, the sample is tilted by 60˚ in the
ψ direction to bring the ZnO(1010) plane normal to the X-ray scattering vector [Figure 2.14
(c)]. If we set θ-2θ value as the Bragg angle of the ZnO (1010) plane (2θ = 31.8˚), Bragg’s
diffraction condition is satisfied in this situation. If we measure X-ray diffraction pattern
in the range of φ between 0˚ and 360˚ while keeping 2θ = 31.8˚, we get 2 peaks whose
positions are separated by 180˚ because two sets of 1010planes can satisfy the Bragg
condition during the φ-scan. XRD in-plane scan for Al2O3 (0006) plane was conducted
with a similar manner.
2.4.2 Dependence of crystallinity on deposition temperature
Figure 2.15 shows the XRD rocking curves for ZnO(1120) peaks of ZnO films
deposited at various substrate temperatures. The FWHM’s for these curves are 1˚, 0.5˚
and 0.9˚ for films deposited at 500˚C, 600˚C and 700˚C, respectively. Figure 2.15 shows
the temperature dependence of the FWHM values for both the XRD peaks (θ-2θ scan)
and the rocking curves. The FWHM values for both decrease as temperature increases
from 400 ˚C to 600 ˚C. However, these trends diverge as temperature is increased further
to 700˚C : The FWHM value for the rocking curve increases while the FWHM of the
XRD peak (θ-2θ scan) remains unchanged. The decrease in FWHM of both XRD peaks
and rocking curves is consistent with increasing grain size and improved epitaxial
30
alignment of the films up to 600˚C. However, it appears that this alignment deteriorates
above 600˚C.
Figure 2.14. Illustration of the in-plane φ-scan measurement. (a) An illustration of the
XRD geometry and the definition of the angle θ, φ and ψ. (b) Shows the ZnO film on r-
plane sapphire with (1120) plane parallel to r-plane as determined in θ-2θ scan. (c)
Shows the situation when sample is tilted by 60° along with ψ direction.
60° (c)
2θ
θ
φ
ψ
ZnO(1010) plane
Sample stage
X-ray
Substrate (Al2O3) ZnO film
φ
ψ 2θ
θ
Detector X-ray source
Substrate with film Sample stage (a)
(b)
2θ
θ φ
ψ
Sample stage
Substrate (Al2O3)
ZnO(1120) plane
X-ray
ZnO film
31
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.00.0
0.2
0.4
0.6
0.8
1.0 700°C 600°C 500°C
Rel
ativ
e in
tens
ity (a
.u.)
ω (deg)
Figure 2.15. XRD rocking curves for ZnO(1120) peak of ZnO films deposited at 500˚C,
600°C and 700˚C.
400 500 600 7000.0
0.5
1.0
1.5
2.0
2.5
Δ2θ
or Δ
ω (d
egre
es)
Temperature (°C)
FWHM (Δω) FWHM (Δ2θ)
Figure 2.16. Temperature dependence of FWHM values of XRD rocking curves (∆ω)
and XRD diffraction peaks in the θ-2θ scan (∆2θ).
32
2.4.3 Surface morphology
The SEM images of ZnO films deposited at 500 ˚C, 600 ˚C and 700 ˚C are shown
in Figure 2.17. SEM images reveal that the film deposited at 500 ˚C has very smooth
surface [Figure 2.17 (a)] but as the deposition temperature increases, poles growing out
of the substrate plane appear. Figure 2.17 (b) shows a few poles on the surface of the film
deposited at 600 ˚C. The density of poles is much larger on the film deposited at 700 ˚C
[Figure 2.17 (c)].
Figure 2.17. SEM images of ZnO films deposited at (a) 500 ˚C, (b) 600 ˚C and (c) 700
˚C. SEM in (a) is the cross section image so that bottom is sapphire substrate and upper is
film. SEMs shown in (b) and (c) are the top views. During deposition, Ar flow rate, O2
flow rate and pressure are kept at 100 sccm, 300 sccm and 5.4 Torr, respectively. The
scale bars are 1 µm.
(a) (b) (c)
33
Figure 2.18 shows the AFM images of the ZnO film surfaces. The root-mean-
square surface roughness values are 0.394 nm, 0.391 nm and 5.49 nm over 1 µm × 1 µm
area for 400˚C, 500˚C, and 600˚C, respectively. The surfaces are very smooth at 400˚C
and 500˚C while the surface of the sample grown at 600˚C has grooves and ridges
aligned in a specific direction. It was not possible to do AFM on the 700 ˚C as the surface
was extremely rough with poles sticking off the surface. Xie et al. also observed such
grooves and ridges parallel to [1101] direction of the sapphire substrate.13 I believe that
this morphology is developed by anisotropic growth of epitaxial ZnO islands on r-plane
sapphire. ZnO grows much faster along the c-axis than [1010] direction.7 Since the c-axis
of ZnO is parallel to the r-plane sapphire surface, after the nucleation of a ZnO island on
the substrate surface, the islands grow anisotropically and give rise to this ridge
morphology.
In addition to the ridge morphology, the films deposited above 600°C have poles
growing out of the film surface. I think these poles are the result of strain relief. The
lattice mismatch between the ZnO film and the r-plane sapphire is 1.53 % along the c-
axis and 18.3 % perpendicular to the c-axis (Figure 2.13) and these mismatches cause
strain in the deposited films. Figure 2.19 shows the lattice parameter of the ZnO films as
a function of the deposition temperature. The lattice parameter approaches to the value
for bulk ZnO (3.2496 Å)14 as temperature increases. Although the films deposited at low
temperature (below 500˚C) have smooth surface they are severely strained. The strain is
reduced and poles appear as the deposition temperature increases.
34
Figure 2.18. AFM images of ZnO films deposited (a) at 400˚C, (b) at 500˚C and (c) at
600˚C. The scan area is 1 µm × 1 µm.
400 500 600 7003.248
3.252
3.256
3.260
ZnO
Lat
ticep
aram
eter
, a (Å
)
Temperature (°C)
Figure 2.19. Dependence of the ZnO lattice parameter, a, on the deposition temperature.
Lattice parameters are calculated from the XRD ZnO(1120) peak positions.
(a) (b) (c)3 nm 3 nm 30 nm
0 nm 0 nm 0 nm
Al2O3 [1101]
35
2.4.4 Two-step deposition
For applications requiring smooth epitaxial films, it is desired to suppress the
nucleation and growth of poles shown in Figure 2.17 (c). A possibility is that the poles
nucleate and grow directly from the substrate at high temperatures. If this is the case, it
may be possible to suppress the poles by first growing a film at low temperature and then
continuing the growth at high temperature. To test this hypothesis, I grew films while the
substrate temperature was changed between two values. For the first two hours the
substrate temperature was kept at 500 ˚C. Following the temperature was changed to 600
˚C or 700 ˚C and the deposition was continued for an additional two hours.
Figure 2.20 shows the SEM images of two films. One of the films was grown for
2 hours at 500 ˚C followed by 2 hours at 600 ˚C [Figures 2.20 (a) and (c)]. The other film
was grown for 2 hours at 500 ˚C followed by an additional 2 hours at 700 ˚C [Figures
2.20 (b) and (d)]. Comparison of Figures 2.20 (a) and (b) with Figures 2.17 (b) and (c)
shows that the number density of poles was reduced in both cases compared to films
deposited at 600 ˚C and 700 ˚C. In this two-step deposition process, the smooth film
deposited during the first step acts as a buffer layer between the sapphire substrate and
film deposited during the high temperature (600˚C or 700 ˚C) period. The buffer layer
plays an improtant role in the morphology of films.15 It seems that there is a competetion
between two types of morphologies during ZnO film growth, poles and smooth films.
The films with a buffer layer which is composed of same materials with the film will
have small lattice mismatch between films and buffer layer and will be grown to smooth
films. On the other hand, the films without buffer layer will have large lattice mismatch
36
and this prevents the film growth and increases the number of poles.
Figure 2.20. SEM images of ZnO films deposited at step temperature profile. (a) and (c)
are the top view and cross section images of a film deposited at 500˚C/2hrs - 600˚C/2hrs.
(b) and (d) are the top view and cross section images of a film deposited at 500˚C/2hrs -
700˚C/2hrs. The scale bar is 1 µm.
2.5 Plasma enhanced chemical vapor deposition of ZnO
It has been reported that the morphology of ZnO thin films can be changed and
high quality ZnO thin films can be deposited at low deposition temperature through
plasma enhanced chemical vapor deposition (PECVD).16,17,18,19 It seems that energetic ion
bambardment onto film surface and radicals generated in a plasma play an important role
in determining the morphology of the ZnO films. Herein, I describe the properties of ZnO
thin films deposited by PECVD.
(a) (b)
(c) (d)
37
2.5.1 Morphology of ZnO thin films deposited by PECVD To generate plasma over a substrate, a 3.5 inch diameter ring type electrode was
placed between showerhead gas distributor and the heated substrate stage in the MOCVD
deposition chamber (section 2.2.2). Radio frequency (rf) power was used to maintain the
plasma. The schematic of PECVD system is shown in Figure 2.21 and process conditions
used in this study are summarized in Table 2.2. ZnO films were deposited under various
conditions in the ranges presented in Table 2.2. Films deposited at 600 ˚C with 20 W rf
power plasma maintained at 3.5 Torr while flowing 200 sccm Ar (carrier gas) and 300
sccm O2 resulted in continuous and smooth films. These process conditions were set as
standard and film properties were investigated by varying one of the process parameters
from this standard conditions.
Table 2.2. PECVD process conditions for ZnO thin film.
Process condition Values Standard
Values
Deposition temperature (˚C) 500 ~ 600 600
Precursor sublime temperature (˚C) 100 100
Carrier gas (Ar) flow rate (sccm) 100 ~ 200 200
O2 flow rate (sccm) 300 300
Process pressure (Torr) 1 ~ 10 3.5
Plasma power (W) 20 ~ 40 20
38
Figure 2.21. Schematic of PECVD system for ZnO thin film deposition. Argon and
oxygen plasma is maintained during deposition by radio frequency power.
Figure 2.22 shows the SEM images of ZnO thin films deposited by PECVD on
silicon substrates at 600 ˚C while maintaining the plasma at 3.5 Torr with RF power
ranging from 20 W to 40 W. While typical ZnO thin films deposited on silicon without
plasma exhibit rough mormophology [Figure 2.10 (a)], films deposited by PECVD seem
to have continuous cross section. The films deposited at rf power of 20 W and 30 W look
continuous and have smooth surfaces. However, the film morphology becomes columnar
and surface roughness increases at 40 W. The anisotropic growth of ZnO seems to be
surpressed at low rf power (below 30 W) but growth again becomes directional at high rf
power (40 W).
39
Figure 2.22. SEM images of ZnO thin films deposited by PECVD at 600˚C and 3.5 Torr.
The applied rf powers were (a) 20 W, (b) 30 W and (c) 40 W, respectively. The scale bars
are 500 nm.
Volintiru et al. suggested a tentative model for the ZnO film growth by PECVD.
They suggested that the electrical and structural properties of films evolve as film growth
goes on.17 To investigate the evolution of film morphology with film thickness, different
thickness ZnO films were deposited at the standard process conditions (Table 2.2) by
adjusting the deposition time. The cross sectional SEM images of these films with
different thicknesses are shown in Figure 2.23. After initial film growth [Figure 2.23 (a)],
the film grows in columnar morphology [Figure 2.23 (b)] but then becomes continuous
after its thickness reaches > 300 nm [Figure 2.23 (c) and (d)]. The two film growth
modes, columnar and contimuous morphology, seem to compete each other in ZnO
depostion by PECVD. The growth with columnar structure is dominant below 300 nm
thickness but continuous morphology becomes dominant as the film thickness increases
above 300 nm. Figure 2.24 shows the FWHM of the XRD peaks (∆2θ) and rocking
curves (∆ω). The XRD measurements show that the FWHM of both the XRD peaks and
the rocking curves decrease with increasing film thickness until the films reach ~ 500 nm.
However, further film growth increases the FWHM. The grain size increases and
(a) (b) (c)
40
crystallinity improves during the initial film growth until the film thickness reaches ~ 500
nm but further film growth decreases the grain size and crystallinity worsens.
Figure 2.23. The cross sectional SEM images of ZnO films deposited by PECVD with
different thicknesses. The films were deposited at standard conditions (Table 2.2) for (a)
30 min, (b) 60 min, (c) 90 min and (d) 120 min.
(a)
(b)
(c)
(d)
130 nm
308 nm
519 nm
935 nm
41
To investigate the dependence of the electrical properties on film thickness,
carrier concentration, mobility and resistivity were determined by using Hall-effect
measurements with Van der Pauw configuration. Figure 2.25 shows the mobility, carrier
concentration and resistivity as a function of film thickness. The mobility increases and
carrier concentration decreases with increasing film thickness from ~ 100 nm to ~ 900
nm. The mobility of polycrystalline thin film is affected by carrier scattering mechanisms
such as grain boundary scattering and ionized defect scattering.20,21,22 The increase in
mobility with increasing film thickness may be due to the reduced grain boundary
scattering because of the grain size increase which is indecated by the narrowed FWHM
of XRD peaks [Figure 2.24].
0 200 400 600 800 10000.15
0.20
0.25
0.30 Δ 2θ
FWH
M (d
eg)
Thickness (nm)
2
3
4
5
6
7
FWH
M (d
eg)
Δ 2ω
Figure 2.24. The plot of FWHM of XRD θ-2θ scans and rocking curves of ZnO films versus film thickness. The measurements were conducted on the same films in Figure 2.23.
42
0 200 400 600 800 10000.04
0.08
0.12
0.16
0
2
4
6
8
10
Thickness (nm)
Res
istiv
ity (Ω
-cm
)
(b)Car
rier c
once
ntra
tion
(1018
/cm
3 )
8
12
16
20
24
28
Mob
ility
(cm
2 /V-s
)
(a)
Figure 2.25. (a) Carrier concentration and mobility and (b) resistivity as a function of
ZnO film thickness. Films were deposited by PECVD on silicon under standard process
conditions (Table 2.2). Film thickness was varied by adjusting the deposition time : 30
min, 60 min and 120 min, respectively in the order of thickness.
2.5.2 Effect of deposition pressure on the film morphology and
optical transmittance
Volintiru et al. reported that the pressure plays an important role in determining
the ZnO film morphology in PECVD.17 Specifically, deposition at low pressures
produces films with smooth surfaces and small grains while deposition at high pressures
43
produces films with rough surfaces and large grains. To investigate the effect of pressure
on the film morphology, ZnO films were deposited under standard conditions (Table 2.2)
but at 3.6 Torr and 1.5 Torr. The SEM images and XRD from the films are shown in
Figure 2.26 and Figure 2.27, respectively. The film deposited at high pressure (3.6 Torr)
shows columnar morphology and an intense XRD peak while the film deposited at low
pressure (1.5 Torr) shows a continuous cross section with smooth surface but a very weak
XRD peak. The FWHM of the (0002)ZnO XRD peak increases as deposition pressure
decreases. Although the film deposited at low pressure shows smooth surface and
continuous cross section, the grain size decreases and the film crystallinity becomes
worse.
Optical transmittance of ZnO films were measured by depositing films on glass
substrates. Figure 2.28 shows the effect of deposition pressure on the optical
transmittance of films. The films with smooth surface continuous cross sections
(deposited at 1 Torr and 1.5 Torr) were brown colored and the film transmittances were
significantly reduced in the visible region. The film with columnar morphology
(deposited at 3.7 Torr) is transparent above 380 nm (~ 3.4 eV, ZnO band gap). The
absorption of the ZnO films in the visible region may be due to carbon incorporation into
the growing film at low pressure. The metal-organic precursor may not be completely
oxidized at low deposition pressure. In conclusion, deposition at low pressure yields
continuous and smooth film but the transmittance of the films in the visible region is
severely reduced.
44
Figure 2.26. The SEM images of ZnO thin films deposited by PECVD on silicon. The
films were deposited under standard conditions (Table 2.2) but at (a) 3.6 Torr and (b) 1.5
Torr.
0
400
800
1200
30 32 34 36 38 400
3
6
(a)ZnO(0002)
FWHM : 0.21°
2θ (deg)
Inte
nsity
(10
coun
ts)
(b) ZnO(0002)ZnO(1011)FWHM : 0.48°
Figure 2.27. The XRD of films deposited under standard conditions (Table 2.2) but at (a)
3.6 Torr and (b) 1.5 Torr. The measured films are same as shown in Figure 2.26.
(a) (b)
100 nm 1 µm
45
300 400 500 600 700 800 900 10000
20
40
60
80
100
% T
rans
mitt
ance
Wavelength (nm)
3.7 Torr 1.5 Torr 1 Torr
Figure 2.28. The effect of pressure on the optical transmittance of thin ZnO films
deposited by PECVD on glass.
46
References
1 Klaus Ellmer, J. Phys. D Appl. Phys. 33, R17 (2000).
2 Y. Chen, D. M. Bagnall, H. Koh, K. Park, K. Hiraga, Z. Zhu, T. Yao, J. Appl. Phys. 84,
3912 (1998).
3 A. Ohtomo, A. Tsukazaki, Semicond. Sci. Tech. 20, S1 (2005).
4 B.P. Zhang, N.T. Binh, K. Wakatsuki, N. Usami, Y. Segawa, Appl. Phys. A 78, 25
(2004).
5 B. Cao, W. Cai, H. Zeng, G. Duan, J. Appl. Phys. 99, 073516 (2006).
6 J. B. Mooney, S. B. Radding, Annu. Rev. Mater. Sci. 12, 81 (1982).
7 J. B. Baxter, E. S. Aydil, J. Cryst. Growth 274, 407 (2005).
8 B. S. Li, Y. C. Liu, Z. S. Chu, D. Z. Shen, Y. M. Lu, J. Y. Zhang, X. W. Fan, J. Appl.
Phys. 91, 501 (2002).
9 M. Ohring, Materials Science of Thin Films, Academic press, USA, P. 279, 2002.
10 R. Triboulet, 10th European Workshop on MOVPE, Lecc (Italy) 8-11 June (2003).
11 M. Kim, Y. Hong, J. Yoo, G. Yi, G. Park, K. Kong, H. Chang, Phys. Stat. sol. (RR) 2,
197 (2008).
12 P. Kung, C.J. Sun, A. Saxler, H. Ohsato, M. Razeghi, J. Appl. Phys. 75, 4515 (1994).
13 J. Q. Xie, J. W. Dong, A. Osinsky, P. P. Chow, T. W. Heo, D. P. Norton, S. J. Pearton,
X. Y. Dong, C. Adelmann, C. J. Palmstrom, Mater. Res. Soc. Symp. P. 891, 407
(2006).
47
14 H. Karzel, W. Potzel, M. Köfferlein, W. Schiessl, M. Steiner, U. Hiller, G. M. Kalvius,
D. W. Mitchell, T. P. Das, P. Blaha, K. Schwarz, M. P. Pasternak, Phys. Rev. B 53,
11425 (1996).
15 D. C. Kim, B. H. Kong, H. K. Cho, D. J. Park, J. Y. Lee, Nanotechnology 18, 015603
(2007).
16 J. H. Park, S. J. Jang, S. S. Kim, B. T. Lee, Appl. Phys. Lett. 89, 121108 (2006).
17 I. Volintiru, M. Creatore, B. J. Kniknie, C. I. M. A. Spee, M. C. M. Sanden, J. Appl.
Phys. 102, 043709 (2007).
18 B. S. Li, Y. C. Liu, D. Z. Shen, Y. M. Lu, J. Y. Zhang, X. G. Kong, X. W. Fan, Z. Z.
Zhi, J. Vac. Sci. Technol. A 20, 265 (2002).
19 Y. J. Kim, H. J. Kim, Mater. Lett. 21, 351 (1994).
20 T. Minami, MRS Bull. 25, 38 (2000).
21 K. Ellmer, G. Vollweiler, Thin Solid Films 496, 104 (2006).
22 M Chen, Z L Pei, X Wang, Y H Yu, C Sun, L S Wen, J. Phys. D Appl. Phys. 33, 2538
(2000).
48
Chapter 3
Properties of ZnO thin films deposited by radio frequency magnetron sputtering
3.1 Introduction
Magnetron sputtering is a physical vapor deposition (PVD) method which makes
use of the ions created in a plasma to transfer materials from a source called target to a
substrate. Ions are accelerated towards to target by an electric field and sputter material
by impinging on the target. Ejected material is deposited on substrates placed across from
the target. A major difference between magnetron sputtering and other thin film
deposition techniques such as thermal evaporation and chemical vapor deposition is that
much higher energy can be input into the growing films with low substrate temperature
because high energy ions in plasma also impinge on the growing film. Thus, the
deposition of high quality films with low substrate temperature is possible.1,2 In addition
to the low deposition temperature, magnetron sputtering has advantage on alloy and
compound deposition by making use of compound target or co-sputtering and it is easy to
scale up to the large size substrate.
Metal oxide thin film deposition by magnetron sputtering can be achieved in two
ways. First, a metal oxide target same as the material to be deposited is used. Second
49
method is called reactive sputtering wherein a metal target is sputtered in the presence of
a reactive gas such as oxygen.3,4 Metal oxide target is usually an insulator and sputtering
requires radio frequency (rf) power. While this lowers the deposition rate, the process is
simple and easy. Reactive sputtering allows one to vary the electrical and optical
properties of the deposited films in a wide range by adjusting the process conditions5,6
such as the reactive gas partial pressure, substrate to target distance and applied power.
In this chapter, the structure and optical and electrical properties of ZnO thin films
deposited by rf magnetron sputtering are described. The dependence of film properties on
key process parameters such as oxygen to argon gas flow rate ratio and deposition
temperature is discussed and the effects of post treatments such as thermal annealing and
oxygen plasma exposure on the film properties are also presented.
3.2 Radio frequency magnetron sputtering for ZnO thin film
Zinc oxide thin films were deposited by rf magnetron sputtering system (AJA
international inc., ATC 2000) from an undoped ZnO target (Williams advanced materials,
ZnO 99.995%) on soda lime glass substrates (microscope slides). The bell jar type
deposition chamber is evacuated by a cryogenic pump which provides a base pressure of
~ 10-7 Torr. The sputtering system has 4 shuttered guns (2 dc-powered guns and 2 rf-
powered guns). A 3-inch-diameter ZnO target is equipped to one of the rf guns. The 3-
inch-diameter target is positioned off the chamber center axis to accommodate four guns
and the target face is tilted toward the sample stage as illustrated in Figure 3.1. To
improve the uniformity of the film thickness, the sample stage is rotated during the
50
deposition. Three process gases (Ar, O2 and N2) can be introduced into the deposition
chamber through separate lines and the flow rates are controlled independently by mass
flow controller (MFC). During sputtering Ar plasma was maintained with 200 W of rf-
power at 5 mTorr pressure. The substrate was heated with halogen lamps located at
beneath the substrate holder and the temperature was controlled by a closed-loop
feedback controller in conjunction with a K-type thermocouple. The sputtering system is
computer operated and process recipes could be programmed and executed automatically.
The diagram of the sputter deposition system is showed in Figure 3.1 and the process
conditions used in this study are listed in Table 3.1.
Table 3.1. Sputtering process conditions for ZnO thin film. The target is undoped ZnO
and substrates are soda lime glasses (microscope slides).
Process factor values
rf power (W) 200
Deposition temperature (˚C) Room temperature, 350
Ar / O2 flow rate (sccm) 20/0, 19/1, 17/3, 15/5
Process pressure (mTorr) 5
51
Figure 3.1. A schematic of the magnetron sputtering system for ZnO thin film
deposition.
3.3 The structure and optical properties of sputtered ZnO films
Figure 3.2 and 3.3 show a typical XRD and an SEM cross sectional image of a
ZnO film sputtered onto a glass substrate. The films deposited under different process
conditions within the process parameter range shown in Table 3.1 resulted films with
similar XRD and looked the same under SEM. Specifically the XRD shows only the
(0002)ZnO diffraction indicating a textured structure with c-axis perpendicular to the
substrate surface. SEM images support this conclusion and show a columnar film
morphology with smooth surface.
52
30 40 50 60 700
1
2
3
4
Inte
nsity
(102 c
ount
s)
2θ (deg)
(0002)ZnO
Figure 3.2. XRD from a ZnO thin film sputtered at room temperature on glass substrate.
The argon and oxygen flow rates were 19 and 1 sccm, respectively.
Figure 3.3. SEM image of a ZnO film sputtered onto a glass substrate at room
temperature. The argon and oxygen flow rates were 19 and 1 sccm, respectively and the
deposition time was 7560 seconds.
Figure 3.4 shows the optical transmittance of three sputtered ZnO films. The three
films had different thickness which was varied by adjusting the deposition time.
500 nm
53
Sputtering for 1680 s, 3360 s and 7560 s resulted in films that were 133 nm, 277 nm and
386 nm, respectively. The average sputtering rate was ~ 5 nm/min during the first ~ 10
minutes but decreases for longer deposition time to ~ 3 nm/min. The transmittance is
almost zero in the ultra violet (UV) region and increases abruptly at around 380 nm
indicating the optical band gap of ~ 3.4 eV. The flims are transparent in visible region for
wavelength > 400 nm. The 133 nm thick flim is still transparent below 380 nm because
the film is still to thin to absorb all the incident light. The interference fringes apprear in
thicker films (277 nm and 386 nm) above 380 nm and indicate that the films have smooth
surface. The spacing of the fringes are consistent with the film thickness which was
determinted by SEM.7
300 400 500 600 700 800 900 10000
20
40
60
80
100
% T
rans
mitt
ance
Wavelength (nm)
133 nm 277 nm 386 nm
Figure 3.4. The transmittance of ZnO films of varying thickness sputtered at room
temperature on glass. The argon and oxygen flow rates were 19 sccm and 1 sccm,
respectively.
54
The ratio of oxygen-to-argon gas flow rate is known as one of the critical process
parameters in oxide thin film deposition.8,9 Accordingly we investigated the dependence
of the film structure and optical properties on the ratio of oxygen-to-argon gas flow rates.
ZnO films were sputtered under various gas flow rate ratios (Table 3.1). The films were
deposited on glass substrate at room temperature for 1680 sec at an average rate of ~ 4
nm/min. Figure 3.5 shows the lattice parameter, c and grain sizes of the ZnO films
deposited at various gas flow rate ratios. The lattice parameters are calculated from the
(0002)ZnO XRD peak position and grain sizes were estimated from the FWHM of the
same peaks using Debye-Scherrer formula. The lattice parameters for all the films were
around 5.250 ± 0.003 Å which is different from the equilibrium value, 5.2038Å.10 This
lattice parameter offset from the equilibrium value may be due to stress in the films.
Assuming that the volume of unit cell is the same, the larger lattice parameter, c, implies
that the lattice is compressed in the a-direction and the films sputtered at room
temperature seem to be under compression parallel to the a-direction.
Figure 3.6 shows the deposition rate as a function of O2 concentration in the feed
gas. The deposition rate decreases as the concentration of oxygen in the feed gas
increases and this is a typical trend of reactive sputtering with O2. In reactive sputtering
from a metal target, the taget surface remains metallic under low oxygen partial pressure
and the deposition rate remains high because of the high sputtering yield of metals.
However, as the oxygen partial pressure increases the target surface oxidizes which
reduces the sputtering and, hence, the deposition rate.1,11
55
0 5 10 15 20 255.20
5.21
5.22
5.23
5.24
5.25
5.26
Latti
ce c
onst
ant (Å)
% O2 in feed gas (O2/(Ar+O2))
30
35
40
45
50
Gra
in s
ize
(nm
)
Figure 3.5. The plot of lattice constant, c, and grain sizes of ZnO films versus the
concentration of O2 in feed gas. The grain sizes are calculated by Sherrer’s formula from
FWHM’s of (0002)ZnO peaks.
0 5 10 15 20 253
4
5
6
Dep
ositi
on ra
te (n
m/m
in)
% O2 in feed gas (O2/(Ar+O2))
Figure 3.6. The plot of film deposition rate versus the concentration of O2 in feed gas.
56
3.4 Effect of post sputtering annealing on the structure and
electrical properties of ZnO thin films The fact that deposition temperature and oxygen partial pressure during
deposition have an important role on the structure and electrical properties of sputtered
ZnO thin films12,13,14 hints that the film quality can be improved by post-deposition
annealing. Two kinds of post deposition annealing were tried here, thermal annealing and
oxygen-plasma annealing.15,16 To investigate the effect of these post-deposition annealing
treatments on the films’ electrical and structural properties, ZnO films were sputtered at
room temperature while varying the concentration of O2 in the feed gas. Following, two
types of annealing were implemented. One set of films were thermal annealed at 600˚C
for 2 hours in vacuum. A second set of films were exposed to oxygen plasma at 600˚C for
2 hours. For both annealing treatments, the PECVD chamber described in section 2.5 was
used. Oxygen plasma was generated at ~ 4.5 Torr with 40 W rf-power applied through
the ring type electrode located 0.5 inch above sample surface. Oxygen gas flow rate was
400 sccm.
Figure 3.7 shows the lattice parameter, c, and the grain size of post treated ZnO
films extracted from XRD (0002)ZnO peak as a function of oxygen concentration in the
feed gas for ZnO sputtering after thermal annealing with and without oxygen plasma. The
lattice parameters of the as-deposited films are far above the relaxed value of 5.2038 Å
indicating that these films are stressed. The stress seems to be relieved after the annealing
steps : the lattice parameter, c, relaxes to 5.2 Å. The relaxation seem to be due to thermal
annealing since annealing without oxygen plasma also relaxes the stress in the films
57
[Figure 3.7 (a)]. The grain sizes increase by both thermal annealing with and without
oxygen plasma as shown in Figure 3.7 (b).
5.20
5.21
5.22
5.23
5.24
5.25
5.26
0 5 10 15 20 2530
40
50
60
Latti
ce p
aram
eter
, c (Å
)
Thermal + plasma Thermal As deposited
(a)
Gra
in s
ize
(nm
)
% O2 in feed gas (O2/(Ar+O2))
Thermal + plasma Thermal As deposited
(b)
Figure 3.7. The plot of (a) lattice constant “c” and (b) grain sizes of as-deposited, thermal
annealing treated and oxygen plasma with thermal annealing treated ZnO thin films
versus the concentration of O2 in feed gas.
Figure 3.8 shows the effect of the thermal annealing in presence of oxygen plasma
on the lattice parameter, c, and grain size as a function of the film thickness. The lattice
parameter, c, for as-deposited films decreases with increasing film thickness indicating
58
within the films, stress relaxes as the films get thicker. After annealing, the lattice
parameter of the films approaches to the equilibrium value indicating that the film
stresses are relaxed upon thermal annealing in presence of an oxygen plasma. This stress
relief is much more effective on thin films (133 nm). Thicker films still exhibit lattice
parameter values larger than the equilibrium value even after annealing. Annealing also
increases the grain size in films as shown in Figure 3.8 (b).
5.21
5.22
5.23
5.24
5.25
100 150 200 250 300 350 40030
36
42
48
54
60
Latti
ce p
aram
eter
, c (Å
)
As deposited 2 hour annealing 5 hour annealing
(a)
Gra
in s
ize
(nm
)
Thickness (nm)
As deposited 2 hour annealing 5 hour annealing
(b)
Figure 3.8. (a) Lattice parameter, c and (b) grain size of as-deposited ZnO films before
and after 2 hour and 5 hour thermal annealing in presence of an oxygen plasma as a
function of film thickness.
59
To investigate the dependence of film structure on thermal annealing temperature,
~ 100 nm thick ZnO thin films were sputtered at room temperature for 1680 sec and then
thermally annealed in vacuum for 1 hour at different temperatures (150, 250, 350 and
450˚C). Figure 3.9 shows the lattice parameter, c, and grain size within these films as a
function of the annealing temperature.
0 100 200 300 400 50030
35
40
45
50
55
60
Gra
in s
ize
(nm
)
Annealing temperature (°C)
5.21
5.22
5.23
5.24
5.25
Lat
tice
cons
tant
(Å)
Figure 3.9. The grain size and lattice parameter, c, of ZnO films as a function of
annealing temperature. The films were deposited on glass substrates at room temperature
while flowing argon and oxygen at 19 sccm and 1sccm, respectively. Annealing was
conducted under vacuum for 1 hour.
The lattice parameter of the films annealed at room temperature and 150˚C were
approximately the same and greater than the equilibrium value. The lattice parameter
decreased with annealing temperature above 150˚C and then saturated at approximately
60
350˚C. The grain size within these films increased monotonically with annealing
temperature. The fact that the grain size and lattice parameters are similar for the films
deposited at room temperature and annealed at 150˚C may be due to inevitable heating of
the substrate during sputtering at room temperature : The substrate temperature is likely
higher than room temperature and may be closer to 150˚C.
The surface roughness and electrical resistivity of as-deposited and thermally
annealed films, both with and without oxygen plasma, were measured and are listed in
Table 3.2. The films were deposited at room temperature using 19 sccm Ar and 1 sccm
O2 in feed gas. The root mean square (RMS) surface roughness was measured using
atomic force microscopy (AFM) in contact mode over a 3 µm × 3 µm area. The electrical
resistivity of the films was measured using contacts in the Van der Pauw configuration
with current-voltage source and measure unit (Keithley 2400).
The RMS roughness of the as-deposited films was ~ 3 nm. Thermal annealing in
absence of oxygen plasma had no effect on the value but when thermal annealing was
done in presence of an oxygen plasma the roughness increased dramatically to ~ 8 nm.
The electrical resistivity of the as-deposited films were above the limit of the
measurement system (~ 107 Ω-cm) and thermal annealing with or without oxygen plasma
did not decrease the resistivity below this limit. Conductive ZnO films are obtained only
when films are sputtered at high temperature. For example, sputtering at 350˚C reduces
the resistivity by 4 orders of magnitude. These results were independent of film thickness
in the range ~ 100 nm - 400 nm.
61
Table 3.2. Surface roughness and electrical resistivity of as-deposited and thermally
annealed films. For comparison, the values for a film deposited at 350˚C are also
presented. Film thickness is the sample thicknesses which are used in resistivity
measurement.
Sample RMS roughness (nm)
Resistivity (Ω-cm)
Film thickness (nm)
As deposited at RT 2.7 > 107 100/125/133/143
Thermal annealing (at 600˚C for 2hrs) 3.9 ~ 107 133/277/386
Oxygen plasma with thermal annealing (at 600˚C for 2hrs)
7.9 > 107 133
As deposited at 350˚C 7.7 4 × 103 544
62
References
1 K. Ellmer, J. Phys. D Appl. Phys. 33, R17 (2000).
2 M. Ohring, Materials Science of thin films, Academic press, USA, 2002.
3 Q. Ma, Z. Ye, H. He, L. Zhu, J. Wang, B. Zhao, Mater. Lett. 61, 2460 (2007).
4 S. B. Krupanidhi, M. Sayer, J. Appl. Phys. 56, 3308 (1984).
5 V. Tvarozek, K. Shtereva, I. Novotny, J. Kovac, P. Sutta, R. Srnanek, A. Vincze,
Vacuum 82, 166 (2008).
6 M.Chen, Z. L. Pei, X. Wang, C. Sun, L. S. Wen, J. Vac. Sci. Technol. A 19, 963 (2001).
7 R. Swanepoel, J. Phys. E Sci. Instrum. 16, 1214 (1983).
8 M. Chen, Z. L. Pei, C. Sun, J. Gong, R. F. Huang, L. S. Wen, Mater. Sci. Eng. B85, 212
(2001).
9 S. Brehme, F. Fenske, W. Fuhs, E. Nebauer, M. Poschenrieder, B. Selle, I. Sieber, Thin
Solid Films 342, 167 (1999).
10 R. T. Downs, K. L. Bartelmehs, G. V. Gibbs, Am. Mineral. 78, 1104 (1993).
11 Z. Li, W. Gao, Mater. Lett. 58, 1363 (2004).
12 E. Mirica, G. Kowach, P. Evans, H. Du, Cryst. Growth Des. 4, 147 (2004).
13 N. Tsuji, H. Komiyama, K. Tanaka, J. J. Appl. Phys. 29, 835 (1990).
14 G. Xiong, J. Wilkinson, B. Mischuck, S. Tuzemen, K. B. Ucer, R. T. Williams, Appl.
Phys. Lett. 80, 1195 (2002).
15 M. Liu, H. K. Kim, Appl. Phys. Lett. 84, 173 (2004).
16 B. Angadi, H. C. Park, H. W. Choi, J. W. Choi, W. K. Choi, J. Phys. D Appl. Phys. 40,
1422 (2007).
63
Chapter 4
Properties of Cu2O thin film deposited on ZnO by metal organic chemical vapor deposition*
4.1. Introduction
Cuprous oxide (Cu2O) is one of the oldest known and studied semiconductors and
may find applications in light emitting diodes1, spintronics,2 catalysis3 and solar cells.4,5,6
Cu2O has been considered as a potential light absorber in solar cells4,5,6 because it has
high optical absorption coefficient,7 high room temperature mobility and long minority
carrier diffusion length.8 Cu2O thin films grown by methods such as thermal oxidization
of copper9, sputtering10 and electrodeposition11, have been used make solar cells.12,13,14
The first Cu2O solar cells were Schottky diodes based on metal-Cu2O junctions but these
suffered from high density of interfacial states and consequently poor photovoltaic
properties. A homo p-n junction could not be constructed because as deposited Cu2O
films are always p-type and a reproducible n-type Cu2O could not be made.15
Recently, solar cells based on metal oxides and in particular, heterojunctions of
Cu2O with other n-type oxides are attracting attention because of their abundance and
nontoxicity.16 ZnO-Cu2O heterojunction may be particularly attractive because, unlike
* Portions of this chapter were published in: S. Jeong, E. S. Aydil, J. Cryst. Growth 311, 4188 (2009)
64
Cu2O, as-deposited ZnO films are always n-type.17,18 Although the theoretical power
conversion efficiency limit based on the Shockley-Queissier limit is ~ 20 %,19 the
reported efficiencies have not yet exceeded 2 %. Poor quality of films and the resulting
heterointerface have been blamed for the low efficiency and these properties depend
critically on the deposition methods and conditions. ZnO-Cu2O heterojunction may also
find applications in optoelectronic devices such as light emitting diodes.
In this chapter, the properties of Cu2O thin films deposited by metal organic
chemical vapor deposition (MOCVD) on ZnO coated glass substrates are described.
Using X-ray diffraction (XRD) and Hall-effect measurements, the dependence of
structural and electrical properties of films on substrate temperature and film thickness
was investigated. Especially it is shown that Cu2O thin films grown by MOCVD on ZnO
have heteroepitaxy, (220)Cu2O || (0002)ZnO; [001]Cu2O || [1210]ZnO, in spite of the fact that
Cu2O and ZnO have different crystal structures, cubic and hexagonal.
4.2. The structure and electrical properties of Cu2O films on
ZnO
4.2.1. Experiment
ZnO coated glass substrates were prepared by radio frequency (rf) magnetron
sputtering from undoped ZnO target at room temperature. Microscope slide glasses were
65
used as substrates. During sputtering, the argon plasma was maintained using 200 W rf
power at 5 mTorr. The ZnO film thickness varied from 100 to 300 nm by changing the
deposition time; the sputtering rate was ~ 3.6 nm/min. Cu2O films were deposited on
these ZnO-coated glass substrates using the MOCVD system shown in Figure 4.1 from
copper(II) hexafluoroacetylacetonate [Cu(C5HF6O2)2, Cu(hfac)2]. Solid Cu(hfac)2 powder
was placed in a small cylindrical copper container and heated up to 65˚C. The sublimed
Cu(hfac)2 vapor was delivered by flowing 5 sccm (standard cubic centimeter per minute)
argon gas through this Cu(hfac)2 container. The Cu(hfac)2 vapor with Ar carrier gas
flowed into the reaction chamber through a 3”-diameter showerhead type gas distributor
placed 2.5 cm across from substrate. Both oxygen gas (300 sccm) and water vapor (~ 120
sccm) were also introduced into the chamber as oxidants but through separate inlets from
the Cu(hfac)2 to prevent premature homogeneous gas phase reactions in the feed lines.
Argon (carrier gas) and oxygen flows were controlled by mass flow controller (MFC)
while water vapor flow rate was controlled using a needle valve. The water vapor flow
rate was estimated from the pressure increase in the chamber. Water vapor was necessary
to get high deposition rates. No deposition was observed even after 60 minutes without
water. Cu(hfac)2 and water vapor feed lines were heated to prevent condensation.
Substrate was placed on top of a heated stage whose temperature was controlled using a
feedback temperature controller. During deposition the total chamber pressure was
controlled by butterfly valve in front of the pump.
66
Figure 4.1. Schematic of the MOCVD system for Cu2O deposition.
Crystal structure of the deposited films were investigated using XRD and in
particular diffraction of CuKα1 radiation from the films. Film morphology was examined
using scanning electron microscopy (SEM). Carrier density and mobility were
determined from room temperature Hall-effect measurements using the Van-der-Pauw
configuration and gold contacts. Optical absorption was measured in transmission using a
combination of deuterium and tungsten halogen lamps (DH–2000–BALL) and a
spectrometer (Ocean Optics HR2000) in the 200 – 1100 nm range.
67
4.2.2. Results and discussion
ZnO thin films sputtered on amorphous glass substrates show textured structure
with (0002) direction parallel to substrate normal. XRD from Cu2O films deposited on
these substrates by MOCVD show only the Cu2O phase without any detectable CuO and
the films have preferred orientation with (220)Cu2O || (0002)ZnO [Figure 4.2 (a)]. If the
ZnO film surface was rough the deposited Cu2O also showed the orientation of (200)Cu2O
|| (0002)ZnO. It was discovered that smooth ZnO surfaces with root mean square (RMS)
roughness smaller than 4 nm across a 3 µm × 3 µm area consistently lead to Cu2O films
with (220)Cu2O || (0002)ZnO [Figure 4.2(a)] whereas Cu2O films deposited on rougher ZnO
surfaces grew with (200)Cu2O || (0002)ZnO orientation [Figure 4.2 (b)]. Examination of the
films with SEM showed that the Cu2O films are facetted [Figure 4.2 (c)] and consisted
with a polycrystalline film.
No film deposition was observed at or below a substrate temperature of 250˚C but
deposition rate increased to 2.2, 4.6 and 14.3 nm/min at 300˚C, 350 and 400˚C,
respectively with pressure kept at 2 Torr. The activation energy of the reaction that limits
the deposition rate was estimated from an Arrhenius’ plot of the deposition rate vs. 1/T
and was found to be 60 kJ/mol (Figure 4.3). This value is close to the apparent activation
energy of Cu2O deposition from copper acetylacetonate [Cu(C5H7O2)2], a precursor
nearly identical to Cu(hfac)2 wherein the F atoms are replaced by H; Condorelli et al.
found this activation energy to be 48 kJ/mole and interpreted it as the activation energy
for dissociative adsorption of Cu(C5H7O2)2.20
68
30 40 50 60 700
2
4
0
30
60
2θ (degrees)
Inte
nsity
(102 c
ount
s)
(0002)ZnO
(200)Cu2O
(110)Cu2O
(0002)ZnO
(220)Cu2O
Figure 4.2. (a) and (b) are typical XRD from Cu2O thin films deposited at 400˚C on ZnO
film. (a) shows XRD of a Cu2O film deposited on a smooth ZnO surface. The RMS
roughness was 3.5 nm and (b) shows XRD from a Cu2O film deposited on a rough ZnO
surface with RMS roughness of 7.7 nm. (c) Typical SEM image of a Cu2O film deposited
by MOCVD at 400˚C. This film was deposited on ZnO coated glass substrate at a total
pressure of 2 Torr for 28 minutes.
(a)
(b)
(c)
100 nm
69
1.5 1.6 1.7 1.81
10
Dep
ositi
on ra
te (n
m/m
in)
1000/T(K-1)
680 660 640 620 600 580 560 Temperature (K)
Figure 4.3. Arrhenius plot of the Cu2O deposition rate at a deposition pressure of 2 Torr.
The linear line is the fitted line.
0.5 1.0 1.5 2.0
5
10
15 350°C 400°C
Dep
ositi
on ra
te (n
m/m
in)
Pressure (Torr)
Figure 4.4. The plot of deposition rate versus process pressure. The films were deposited
at two substrate temperature, 350 and 400˚C.
70
If the deposition is limited by dissociative adsorption of Cu(hfac)2 the deposition
rate should also increase with increasing partial pressure of Cu(hfac)2. We increased the
partial pressure of Cu(hfac)2 by reducing the pumping speed with a butterfly valve while
keeping the flow rates of all gases constant; this increases the deposition pressure and
consequently the partial pressure of all gases including the precursor Cu(hfac)2. Indeed,
the deposition rate increased with pressure at constant temperature as shown in Figure
4.4.
Figure 4.5 shows the morphology and optical absorbance of the Cu2O films
deposited at temperatures ranging from 300˚C to 400˚C. Deposition durations were
adjusted such that all the films in Figure 4.5 have similar thicknesses: films deposited at
300, 350 and 400˚C were 222 nm, 208 nm, and 243 nm thick, respectively. Cross
sectional SEM images [Figure 4.5 (d) - (f)] show that the films deposited at low
temperature (300˚C) are rough but become denser and smoother with increasing substrate
temperature. The rough surface at low temperature may be a result of slow surface
diffusion of the precursor adsorbate that eventually leads to Cu2O. The rough surface
affects the optical appearance of the film. The films deposited at 350˚C and 400˚C are
semitransparent in the visible [Figure 4.5 (b) and (c)] but the film deposited at 300˚C
[Figure 4.5 (a)] is opaque due to scattering. Even though all the films are similar in
thickness, the film deposited at low temperature (300˚C) looks optically thicker [Figure
4.5 (a) - (c)] because of light scattering by the rough surface. Figure 4.5 (g) shows the
optical absorbance of the films in Figure 4.5 (a) - (f). The films deposited at 350˚C and
400˚C start to absorb light at around 600 nm.
71
400 600 800 10000
1
2
3
Abs
orba
nce
Wavelength (nm)
300°C 350°C 400°C
Figure 4.5. (a) - (c) Digital photographs and (d) - (f) cross sectional SEM images of
Cu2O films deposited at substrate temperatures of 300 (a, d), 350 (b, e) and 400˚C (c, f),
respectively. Scale bars are 500 nm. (g) Optical absorbance of the films shown in (a) - (f).
(g)
(a)
(b)
(c) (f)
(e)
(d)300˚C
350˚C
400˚C
300˚C
350˚C
400˚C
72
The Tauc plot (Fig. 4.6) generated from the absorption data in Figure 4.5 (g)
indicates that the optical band gap of the deposited films are around 2 - 2.2 eV near the
value expected for Cu2O which has a band gap of 2.17 eV. The absorption curve in the
2.0 - 2.2 eV range seems to have fine structure which may be resolved if the absorption
measurements are done at low temperature but are smeared out at room temperature.
1.5 2.0 2.5 3.00
1x1010
2x1010
3x1010
4x1010
350°C 400°C
(αE)
2 (eV/
cm)2
Energy (eV)
Figure 4.6. Tauc plot generated from the absorption data in Figure 4.5 for films
deposited at 350˚C and 400˚C.
XRD peaks of all the Cu2O films deposited at different substrate temperatures
show only the ZnO (0002) and Cu2O (220) diffraction peaks (Figure 4.7) and this
indicates that Cu2O grows on ZnO by MOCVD with preferred orientation and (220)Cu2O ||
(0002)ZnO. This orientation is different than those reported for other deposition methods
73
such as sputtering and electrodeposition, which result in films where (111)Cu2O ||
(0002)ZnO.11,17
0
4
8
0
4
8
30 40 50 60 700
4
8
300°C
(0002)ZnO
(220)Cu2O
Inte
nsiy
(102 c
ount
s)
350°C
2θ (degrees)
400°C
Figure 4.7. XRD peaks from the films deposited at different substrate temperatures 300,
350 and 400˚C. All the films were deposited under a total pressure of 2 Torr and the film
thicknesses were made similar (222 nm, 208 nm and 243 nm respectively) by adjusting
the deposition times.
Cu2O grain size was estimated from the full width half maximum (FWHM) of the
(220) XRD peak. Figure 4.8 (a) shows that the grain size increases with deposition
temperature. However care is require in estimating the grain size from XRD because the
XRD peak width is affected by both grain size and stress. Decreasing XRD peak width
74
may be due to both increasing the grain size and decreasing stress. To investigate the film
stress, Cu2O lattice constant was determined from the location of the (220) diffraction
peak. The equilibrium Cu2O bulk lattice parameter is 4.267Å. It was found that the lattice
parameter of the Cu2O films was closest to this equilibrium value when the films were
deposited at 300oC but the lattice parameter increased with increasing deposition
temperature. At low temperature, the films are made up of small grains and the stress due
to lattice mismatch at the interface is relaxed. At high substrate temperature, grains grow
larger, crystallinity improves and epitaxy yields a smooth and oriented film with
(220)Cu2O || (0002)ZnO. The increase in the intensity of the Cu2O (220) diffraction with
increasing substrate temperature (Figure 4.7) also indicates that high deposition
temperature results in well oriented crystalline films. However this improvement in grain
size and crystallinity increases the stress in the film. Figure 4.8 (a) shows that the lattice
constant increases above the equilibrium value with increasing temperature. The film
deposited at 400˚C shows ~ 0.5 % strain. Thus, the narrowing of the XRD peak is largely
due to the increased grain size and not due to strain relaxation. In fact, increasing strain
likely counteracts the narrowing of the XRD peak so that the grain size estimates are
lower bounds.
Figure 4.8 (b) and (c) show the dependence of the carrier mobility, carrier
concentration and Cu2O film resistivity on the deposition temperature. While carrier
concentration does not vary significantly with the deposition temperature, the carrier
mobility increases by a factor of ten as the deposition temperature is increased from
300˚C to 400˚C. Consequently, the resistivity decreases with increasing deposition
temperature. The carrier mobility in polycrystalline thin films with small grains is usually
75
limited by frequent scattering of the carriers at the grain boundaries. Consistent with this
expectation, the increase in the carrier mobility correlates with increase in grain size and
concomitant reduction in grain boundary scattering at higher deposition temperatures.
20
40
60
4.26
4.27
4.28
4.29
6
12
0
10
20
30
300 350 4000
1
2
3
Gra
in s
ize
(nm
)
(a)
Lat
tice
cons
tant
(Å)
Car
rier c
once
ntra
tion
(1014
/cm
3 )
(b)
Mob
ility
(cm
2 /V s
)
Res
istiv
ity(1
03 Ω c
m)
Temperature (°C)
(c)
Figure 4.8. (a) Grain size and lattice parameter of the Cu2O films as a function of
deposition temperature. Equilibrium lattice constant of Cu2O is 4.267Å. (b) Carrier
concentration, carrier mobility and (c) resistivity as a function of deposition temperature.
It was discovered that the Cu2O film properties depend on the film thickness. The
XRD patterns from different thickness Cu2O films deposited at the same substrate
76
temperature (350˚C) all show the same preferred orientation, (220)Cu2O || (0002)ZnO
(Figure 4.9) but the grain sizes and the Cu2O lattice parameters determined from these
XRD peaks decrease with increasing film thickness [Figure 4.10 (a)]. In particular, the
lattice parameter approaches the equilibrium value for bulk Cu2O as the film thickness
increases. It appears that the stress in the films relaxes with increasing film thickness.
0
3
6
0
3
6
30 40 50 60 700
3
6
224nm(0002)ZnO
(220)Cu2O
Inte
nsity
(102 c
ount
s)
329nm
2θ (degrees)
509nm
Figure 4.9. XRD from Cu2O films deposited at the same temperature (350˚C) as a
function of film thickness.
Generally, grain size is expected to increase with increasing film thickness in
polycrystalline thin films but Cu2O films deposited by MOCVD show the opposite trend.
Thin Cu2O films deposited at high temperature (~ 400˚C) have large grains and the film
77
develops high stress to maintain the epitaxy and the textured structure on ZnO. As films
get thicker this stress is released by forming small grains and the film develops a rough
surface morphology. This morphological change with film thickness has profound effect
on the electrical properties of the film. Specifically, the carrier mobility decreases with
increasing thickness because the grain size decreases [Figure 4.10 (b)]. Consequently, the
resistivity increases with increasing film thickness [Figure 4.10 (c)].
30
40
50
60
4
8
12
200 300 400 5000
3
6
9
Gra
in s
ize
(nm
) (a)
4.26
4.27
4.28
4.29
Lat
tice
cons
tant
(Å)
Car
rier c
once
ntra
tion
(1014
cm
-3)
(b)
10
20
30
Mob
ility
(cm
2 /Vs)
Res
istiv
ity(1
02 Ω c
m)
Thickness (nm)
(c)
Figure 4.10. (a) Grain size and lattice parameter of the Cu2O films as a function of film
thickness. (b) Carrier concentration, carrier mobility and (c) resistivity as a function of
film thickness. Deposition temperature was 350˚C for all films.
78
4.2.3 Conclusion
Cu2O thin films were deposited by MOCVD on ZnO coated glass substrates from
Cu(hfac)2, oxygen gas and water vapor. The films are single Cu2O phase (cubic cuprite
structure) and exhibit preferred growth orientation on ZnO with (220)Cu2O || (0002)ZnO.
The grain size increases with deposition temperature. The lattice parameter of the Cu2O
film also increases with increasing temperature and exceeds the bulk equilibrium value
by as much as 0.5 %. At low deposition temperature (300˚C) the film has rough surface
morphology but as the deposition temperature is increased, the films become smooth with
large grains. However, the film is stressed to maintain epitaxial and textured structure on
ZnO. Larger grains result in higher carrier mobility and consequently resistivity decreases
with increasing deposition temperature. The grain size and lattice parameter of the films
deposited at the same temperature decrease with increasing film thickness indicating that
the film stress relaxes with increasing film thickness. Stress relaxation and smaller grains
with increasing film thickness are accompanied by rough surface morphology and lower
carrier mobility.
4.3 Heteroepitaxy of Cu2O thin film on ZnO
The fact that Cu2O films deposited on ZnO are textured with (220)Cu2O || (0002)ZnO
(Figure 4.7 and 4.9) also raises the possibility that the Cu2O growth on ZnO is epitaxial
with (220)Cu2O || (0002)ZnO. Further experiments discussed below show that this is indeed
the case (vide infra). This textured structure is different from previously reported results
where Cu2O films deposited by sputtering and electrochemical methods showed that
79
(111)Cu2O || (0002)ZnO.11,17 It appears that MOCVD produces a different interface and
epitaxy than these other deposition methods.
The epitaxy between Cu2O and ZnO is surprising because Cu2O and ZnO have
cubic (a = 4.27 Å) and wurtzite (a = 3.25Å and c = 5.21Å) crystal structures, respectively
with large lattice mismatch. However, epitaxy is still possible because these materials
may form an energetically favorable periodic coincidence lattice at the interface between
them. Figures 4.11 (a) and (b) show the atomic arrangements of the Cu atoms on the
Cu2O (220) plane and of the oxygen atoms on the ZnO (0002) plane, respectively. We
can infer from the XRD data in Figure 4.7 and 4.9 that these planes are parallel to each
other and to the glass substrate surface.
Copper atoms on the Cu2O (220) plane and oxygen atoms on the ZnO (0002)
plane form 2 aCu2O × aCu2O (aCu2O = 0.427 nm) and 3 aZnO × aZnO (aZnO = 0.325 nm)
rectangular unit cells, respectively. The longer sides of these unit cells, 2 aCu2O and
3 aZnO, have a lattice mismatch of 7.28 %. However, (aCu2O/aZnO) ≈ (4/3) and epitaxy
with (001) Cu2O || (1210) ZnO would produce a periodically coincident lattice with small
coincidence lattice mismatch, Fo, of 1.46 % [Fo= (4aZnO – 3aCu2O)/4aZnO]. In cases where
afilm/asubstrate = m/n with m and n positive integers and m = n + 1 an edge dislocation is
created to allow lattice matching21 and this would be the case here. Thus, these atomic
arrangements in the Cu2O (220) and ZnO (0002) planes suggest an epitaxial relation
where (001) Cu2O || (1210) ZnO. A proof of this hypothesis is shown in next.
80
Figure 4.11. Atomic arrangements of (a) copper atoms on Cu2O (220) plane and (b)
oxygen atoms on ZnO (0002) plane.
To conclusively determine the epitaxial relationship between Cu2O and ZnO, we
prepared an epitaxial ZnO film on single crystalline sapphire (a-plane) by plasma
enhanced MOCVD (PEMOCVD) from zinc acetylacetonate hydrate [Zn(acac)2] and
oxygen gases. The plasma deposition system was the same as that described in section
2.5. Plasma was generated over the substrate by applying radio frequency (rf) power to
the ring electrode. Zn(acac)2 was heated up to 100˚C in a copper boat and transported to
the reactor by 200 sccm of Ar as carrier gas and 300 sccm of oxygen gas was delivered
through a separate inlet. The substrate was heated up to 600˚C and pressure was kept at
3.6 Torr. The plasma was maintained during deposition by applying 20 W of rf-power to
the ring electrode. Zinc oxide films grow on a-plane sapphire with c-axis perpendicular to
81
the substrate and with in-plane rotational alignment due to [0001]ZnO || [1120] sapphire and
[1120] ZnO || [0001]sapphire epitaxy.22 In this way, we could grow polycrystalline epitaxial
ZnO thin films where the six-fold rotational symmetry of the ZnO crystals around their c-
axis is retained across large areas.
To test the hypothesis for the epitaxial relation (001) Cu2O || (1210) ZnO, a cuprous
oxide film was deposited on this aligned ZnO / Al2O3 substrate. Figure 4.12 shows the
XRD pattern of Cu2O/ZnO heteroepitaxial films grown on Al2O3 substrate and
establishes that Cu2O, ZnO and Al2O3 are aligned such that (220)Cu2O || (0002)ZnO ||
(1120) Al2O3; only the XRD peaks due to these planes are observed. This alignment is
expected from the results presented so far and those in reference 22.
In-plane alignment in Cu2O/ZnO heteroepitaxial films grown on Al2O3 was
investigated by conducting φ-scans of the Cu2O (200) and ZnO (1011) diffractions. These
diffractions are chosen for φ-scans because the zone axis of (220) and (200) planes in
Cu2O is 001⟨ ⟩ [Figure 4.13 (a)] and the zone axis of (0002) and (1011) planes in ZnO is
1210⟨ ⟩ [Figure 4.13 (b)]. For these measurements, we set (200)Cu2O perpendicular to the
X-ray incident plane and fixed 2θ at 42.328˚ [(200)Cu2O diffraction]. The substrate was
then rotated to scan the angle φ across 360˚ (φ-scan). On the same sample, we conduct a
second φ-scan where we set (1011) ZnO perpendicular to the X-ray incident plane and fixed
2θ at 36.29˚ [ (1011) ZnO diffraction].
82
30 40 50 60 70103
105
107
Inte
nsity
(cou
nts)
2θ (degrees)
(110
) Cu2
O (000
2) Z
nO
(112
0) A
l2O3
(220
) Cu2
O
(000
4) Z
nO
Figure 4.12. XRD from a Cu2O film deposited by MOCVD on ZnO which itself was
deposited on a-plane sapphire substrate by PEMOCVD. Cu2O film was deposited at
400˚C and 2 Torr. ZnO film was deposited at 600˚C and 3.6 Torr.
These φ-scans [Figure 4.13 (c) and 4.13 (d)] show six peaks at identical angles for
(200)Cu2O and (1011) ZnO diffractions and conclusively show that (001) Cu2O || (1210) ZnO.
Combining this result with those from 2θ scan, leads to the conclusion that, indeed, Cu2O
films grow on ZnO films epitaxially with (220)Cu2O || (0002)ZnO; (001) Cu2O || (1210) ZnO.
Hence the hypothesis of the heteroepitaxy between Cu2O and ZnO based on the atomic
structure of the Cu2O (220) surface and ZnO (0002) surface is proven to be true. The
sharpness of the peaks and lack of other peaks in Figures 4.13 (c) and 4.13 (d) indicates
nearly perfect alignment.
83
0
1
2
3
-150 -100 -50 0 50 100 1500
2
4
Inte
nsity
(cou
nts)
(c)
φ (degrees)
(d)
Figure 4.13. (a) Cubic Cu2O and (b) hexagonal ZnO unit cells illustrating that the zone
axis of the Cu2O (220) and (200) planes is the 001⟨ ⟩ direction and that the zone axis of
ZnO (0002) and (1011) planes is the 1210⟨ ⟩ direction. XRD φ-scan measurements (c) with
2θ fixed at 42.328˚ [Cu2O (200) diffraction] and (d) at 36.29˚ [ZnO (1011) diffraction].
(a) (b)
84
The six-fold rotational symmetry of the ZnO around the c-axis had been
established before and is expected.22 However, at first glance, six-fold rotational
symmetry in φ-scan from cubic Cu2O may seem surprising. The six peaks in the φ-scan
from Cu2O can be explained by considering the symmetry of the underlying ZnO and the
possible ways in which (220)Cu2O || (0002)ZnO ; (001) Cu2O || (1210) ZnO epitaxy can be
realized on polycrystalline ZnO films. Given a ZnO grain with in-plane allignment with
respect to the sapphire substrate, the Cu2O crystals can grow on this grain by alligning the
rectangular unit cell shown in Figure 4.11(a) on the ZnO (0002) plane in three different
directions separated rotationally by 60˚. These alignments are illustrated in Figure
4.11(b). Since the ZnO film is polycrystalline and a large number of grains are sampled
during diffraction, all the possible orientations shown in Figure 4.11(b) are equally likely
and give rise to the six peaks in the φ-scan.
85
References
1 L. T. Drapak, Semiconductors 2, 624 (1968).
2 S. J. Pearton, W. H. Heo, M. Ivill, D. P. Norton, T. Steiner, Semicond. Sci. Tech. 19,
R59-R74 (2004).
3 J. Ramirez-Ortiz, T. Ogurab, J. Medina-Valtierra, S. E. Acosta-Ortiz, P. Bosch, J. A. de
los Reyes, V. H. Lara, Appl. Surf. Sci. 174, 177-184 (2001).
4 J. Herion, E. A. Niekisch, G. Scharl, Sol. Energ. Mater. 4, 101 (1980).
5 L. Papadimitriou, N. A. Economou, D. Trivich, Sol. Cells 3, 73 (1981).
6 L.C. Olsen, F. W. Addis, W. Miller, Sol. Cells 7, 247 (1982-1983).
7 A. E. Rakhshani, Solid State Electron. 29, 7 (1986).
8 C. A. Dimitriadis, L. Papadimitriou, N. A. Economou, J. Mater. Sci. Lett. 2, 691 (1983).
9 B. P. Rai, Sol. Cells 25, 265 (1988).
10 S. Ishizuka, S. Kato, Y. Okamoto, K. Akimoto, J. Cryst. Growth 237, 616 (2002).
11 M. Izaki, T. Shinagawa, K. Mizuno, Y. Ida, M. Inaba, A. Tasaka, J. Phys. D Appl.
Phys. 40 3326 (2007).
12 A. Mittiga, E. Salza, F. Sarto, M. Tucci, R. Vasanthi, Appl. Phys. Lett. 88, 163502
(2006).
13 T. Minami, T. Miyata, K. Ihara, Y. Minamino, S. Tsukata, Thin Solid Films 494, 47
(2006).
14 S.S. Jeong, A. Mittiga, E. Salza, a. Masci, S. Passerni, Electrochim. Acta. 53, 2226
(2008).
15 H. Raebiger, S. Lany, A. Zunger, Phys. Rev. B 76, 045209 (2007).
86
16 H. Tanaka, T. Shimakawa, T. Miyata, H. Sato, T. Minami, Thin Solid Films 469-470,
80 (2004).
17 K. Akimoto, S. Ishizuka, M. Yanagita, Y. Nawa, G. K. Paul, T. Sakurai, Sol. Energy
80, 715 (2006).
18 S. Ishizuka, K. Suzuki, Y. Okamoto, M. Yanagita, T. Sakurai, K. Akimoto, N.
Fujiwara, H. Kobayashi, K. Matsubara, S. Niki, Phys. Status Solidi C 1, 1067 (2004).
19 W. Shockley, H. J. Queisser, J. Appl. Phys. 32, 510 (1961).
20 G. G. condorelli, G. Malandrino, I. Fragala, Chem. Mater. 6 1861 (1994).
21 A. Trampert, K. H. Ploog, Cryst. Res. Technol. 35, 793 (2000).
22 J. B. Baxter, E. S. Aydil, J. Cryst. Growth 274, 407 (2005).
87
Chapter 5
Cu2O/ZnO heterojunction and photovoltaic application
5.1. Overview
In this chapter, current-voltage (J-V) characteristics of Cu2O/ZnO heterojunction
thin film solar cells are described. Carrier transport and recombination mechanism were
investigated through the analysis of their current-voltage (J-V) characteristics in the dark
and under various illumination intensities as a function of temperature between 100 K
and 300 K. The temperature dependence of diode ideality factor, short circuit current and
open circuit voltage within analytic models of this heterojunction was discussed and
activation energy barriers for current flow were extracted. The Cu2O/ZnO heterojunction
solar cells were prepared by metal organic chemical vapor deposition (MOCVD) of Cu2O
thin films on ZnO sputtered on transparent conducting oxide (TCO) coated glass
substrates. Activation energies extracted from temperature dependence of the saturation
current reveals that interface recombination is the dominant carrier transport mechanism
in these solar cells. Moreover, tunneling across an interfacial barrier plays an important
role in current flow and TiO2 buffer layer reduces tunneling. A high open circuit voltage
at low temperature (~ 0.9 V at around 100 K) indicates that Cu2O/ZnO heterojunction
88
solar cells have high potential as solar cells if the recombination and tunneling at the
interface can be suppressed at room temperature.
5.2 Cu2O/ZnO heterojunction solar cell fabrication
Cu2O/ZnO heterojunction thin film solar cells were prepared on tin doped indium
oxide (ITO) coated glass substrates by successive deposition of Al-doped ZnO (AZO)
and undoped ZnO by magnetron sputtering followed by metal organic chemical vapor
deposition (MOCVD) of Cu2O. Radio frequency (RF) magnetron sputtering of ZnO and
Al-doped ZnO was conducted at room temperature from undoped- and Al-doped-ZnO
targets, respectively. An argon plasma maintained with 200 W of rf power at 5 mTorr
was used for sputtering. The Al-doped ZnO layer thickness was kept constant at 200 nm
but the effect of undoped ZnO thickness was examined by varying it between 65 nm and
200 nm. In one of the solar cells, 10 nm thick TiO2 layer was deposited between the
Cu2O and a 65 nm thick ZnO film to examine the effect of a potential passivating buffer
layer on the charge transport mechanism across the junction. This TiO2 layer was
deposited from Titanium(IV) isopropoxide (Ti[OCH(CH3)2]4) and water vapor using
atomic layer deposition (ALD) at a temperature of 180˚C (Cambridge NanoTech Inc.,
Savannah). The deposition rate was 1.9 Å/cycle.
Cuprous oxide (Cu2O) films were deposited on these glass substrates, now coated
with a stack of ZnO/AZO/ITO or TiO2/ZnO/AZO/ITO, using MOCVD and Copper(II)
hexafluoroacetylacetonate (Cu(hfac)2) as the organometallic precursor. These film stacks
are shown schematically with a real sample picture in Figure 5.1. The detail of the Cu2O
89
MOCVD process was described in chapter 4. Briefly, solid Cu(hfac)2 powder was placed
in an enclosed copper boat heated to 65 ˚C. The Cu(hfac)2 vapor was delivered into the
deposition chamber by flowing argon through the boat as a carrier gas. The substrate was
placed on a platen, which was maintained at 350 ˚C. Both water vapor and O2 gas were
used as oxidants.
Figure 5.1. Schematic of Cu2O/ZnO solar cell structure and a real sample picture. For the
measurement of J-V characteristics, positive bias was applied to the Au pad in the
forward bias case.
90
After the Cu2O thin film deposition, gold pads were deposited by thermal
evaporation for Cu2O-side metal contacts therefore, the two electrodes for cells are ITO
and Au. During the Au evaporation, metal mask with 2 mm diameter holes was used to
define the shape and size of metal pads and this hole’s size is the cell size. Morphology
and thickness of the deposited films were examined using scanning electron microscopy
(SEM). Typical thickness of the Cu2O film in the solar cells discussed herein is 280 nm.
A list of four representative solar cells that will be discussed and compared in this chapter
is given in Table 5.1.
5.3 Solar cell characterization
5.3.1 Characterization background
The performance of solar cells which convert incident photon energy to electrical
energy can be quantitatively analyzed by measuring the current-voltage (J-V)
characteristics under illumination. The J-V characteristics of p-n junction type solar cells
can be described by equivalent electrical circuit model shown in Figure 5.2. In the
simplest case with infinite shunt resistance and zero series resistance, the J-V
characteristics of the solar cell without illumination follow diode curve. If the diode is
ideal and diffusion is the dominant carrier transport mechanism at the junction interface,
its J-V characteristics can be described by the Shockley diode equation,
J = Jo exp qVkT
⎛ ⎝ ⎜
⎞ ⎠ ⎟ −1
⎡
⎣ ⎢
⎤
⎦ ⎥ (5.1)
91
where, q is the electronic charge, kT is the thermal energy and Jo is the reverse saturation
current, respectively. When the solar cell is illuminated, electron-hole pairs are generated
and a photo current flows in the opposite direction of current flow under bias. This photo
generated current shifts the dark J-V characteristics in the negative direction along the
current axis by the amount of the photo current as shown in Figure 5.3. The point where
the J-V characteristic crosses the x-axis (voltage axis) is called the open-circuit voltage
(Voc). This is the voltage of which the current in the external circuit is zero. The point
where J-V curve crosses the y-axis (current-density axis) is called the short-circuit current
(Jsc) which corresponds to the current when voltage is zero. An ideal solar cell will do
work on an external load at a rate equal to the maximum power given by the product |Jsc
× Voc| but a real solar cell does work less than this amount. The real maximum power
(Pmax) will be the value which makes the product |J × V| maximum in the range between 0
and Voc volts (and current density between 0 and Jsc). The ratio of Pmax to |Jsc × Voc| is
called the fill factor (FF) of a solar cell.
Figure 5.2. Schematic of a solar cell equivalent circuit.
92
Figure 5.3. Illustration of a virtual J-V characteristic for a solar cell under illumination.
Vm and Jm are the maximum voltage and maximum current density, respectively.
Maximum power (Pmax) is produced in the ranage 0 < V < Voc and 0 < J < Jsc. Fill factor is
the ratio of |Jm × Vm| to |Jsc × Voc|.
The power conversion efficiency of a solar cell is given by
max sc oc
i i
P FF J VP P
η × ×= = (5.2)
where, Pi is the illumination power density.
To investigate the dependence of conversion efficiency of solar cell on the
incident light wavelength, incident-photon-to-current-conversion-efficiency (IPCE) is
used. The IPCE, which is also referred to as the external-quantum-efficiency (EQE) is
given by
93
currentelectrons (charge of 1 electron)secIPCE =photons (total power of photons)sec (energy of 1 photon)
( ) ( )
= IQE × LHE
i i
I hcIeP e P
hλ λ λ
ν
=
= = (5.3)
where, e is the electronic charge, h is the Planck’s constant, c is the speed of light, λ is the
wavelength and ν is the frequency of the incident light. Therefore, IPCE can be
determined from the measurements of incident power and photo generated current as a
function of incident light wavelength. The IPCE can also be expressed as a product of
internal-quantum-efficiency (IQE) and light-harvesting-efficiency (LHE). IQE is the
measure of the generated carrier transport efficiency to the external load. LHE is
expressed as
LHE 1 10 A−= − (5.4)
where, A is the optical absorbance of the solar cell.
5.3.2 Characterization equipment
The solar cells’ J-V data were collected and recorded using a semiconductor
parameter analyzer (Agilent Technologies, 4155C). For the temperature dependent J-V
characteristics, the data were collected as a function of temperature between 100 K and
300 K in the dark and under illumination in a cryogenic microprobe station (Janis
Research Company ST-500) with optical access through its bottom flange and through
94
the sample platen (Figure 5.4). A Xe-arc lamp in conjunction with a 0.125 m
monochromator was used to illuminate the solar cells. The monochromator is equipped
with both a 600 groove/mm grating and a mirror mounted on the same turret. For this
work, the mirror was selected to illuminate the solar cells with the broadband AM1.5
spectrum (100 mW/cm2). The AM1.5 spectrum was generated using commercially
available filters (Newport Corp.) between the Xe-arc lamp and the monochromator. The
output from the monochromator was directed and imaged onto the plane of the sample
platen using mirrors and quartz lenses. The irradiance, Pi(λ), was recorded at the plane of
the sample platen, where the solar cell is placed, using a calibrated radiometry system
that included a Thermopile sensing probe and a radiant power meter from Newport
(70268 and 70260). The incident power was adjusted in the range from 2 to 200 mW/cm2
using neutral density filters. The temperature of the cryogenic platen was measured at
two locations and controlled using a combination of continuous liquid nitrogen flow and
a heater. However, this temperature is not necessarily the same as the solar cell
temperature due to heat transfer resistance between the solar cell and the cryogenic
platen. Such high heat transfer resistance between the substrate and cooled platen is
common under vacuum. To measure the solar cell temperature, an additional calibrated
temperature sensor (Omega Engineering Inc., Resistance temperature detector) was used.
It was attached directly on the surface of a second solar cell placed next to the solar cell
being tested. The two solar cells were clamped in the same way to the cryogenic chuck so
that monitoring the temperature of the second solar cell gives an accurate measure of the
temperature of the solar cell being studied. All reported temperatures are those measured
by this temperature sensor.
95
Figure 5.4. Equipment for measuring the temperature dependent J-V characteristics of
solar cells. (a) is the photograph of the overall system. (b) is the schematic of the
measurement set up.
(a)
(b)
96
5.4 Photovoltaic characterization for Cu2O/ZnO solar cells
The photovoltaic characteristics of four different solar cells which have the
structure of Au/Cu2O/(TiO2)/ZnO/Al-ZnO/ITO/Glass with different ZnO thicknesses (65
nm, 130 nm, 200 nm and 65 nm with 10 nm TiO2 buffer layer) were investigated in the
dark and under illumination at room temperature. The solar cells without undoped ZnO
layer do not show electrical rectifying characteristics and the fact that sputtered ZnO
layer at room temperature has high resistivity (> 107 Ω-cm, chapter 3) suggests that the
thickness of ZnO has an important role in photovoltaic characteristics. One of the solar
cells has TiO2 layer between Cu2O and ZnO to investigate buffer layer passivation.
The J-V curves for two solar cells with 65 nm and 130 nm thick ZnO in the dark
and under illumination are shown in Figure 5.5. Other solar cells showed similar J-V
characteristics. The figures of merit (Voc, Jsc, FF, efficiency) for these solar cells are listed
in Table 5.1. Several conclusions can be drawn from these J-V characteristics. First, the
J-V characteristics under illumination show that these cells have small Voc (0.19 ~ 0.34 V)
and Jsc (0.64 ~ 0.74 mA/cm2) and as a result, power conversion efficiencies are poor
(below 0.1%). Second, the J-V characteristics of thin ZnO (65 nm) solar cells seem to
follow the exponential diode curve (diode ideality factor is 2.7) but those of the thick
ZnO solar cells (above 130 nm) seem to be limited by series resistance and currents do
not rise exponentially as would be expected from a diode (diode ideality factor is above
7). Third, the curves in the dark and under illumination cross over each other in the
forward bias region and this characteristic is most prominent in solar cells thicker ZnO
films.
97
-0.2 0.0 0.2 0.4 0.6 0.8
-1
0
1
2
3
4
Cur
rent
den
sity
(mA
/cm
2 )
Bias (V)
ZnO 65 nm DarkZnO 65 nm LightZnO 130 nm DarkZnO 130 nm Light
Figure 5.5. J-V characteristics of two Cu2O/ZnO heterojunction solar cells. The cell
structures are Au/Cu2O/ZnO/Al-ZnO/ITO/Glass with 65 nm thick ZnO and 130 nm thick
ZnO. The “Dark” and “Light” J-V characteristics were measured in the dark and the
under illumination, respectively.
Table 5.1. The figures of merit for Cu2O/ZnO heterojunction solar cells. The structure of
solar cells is Au/Cu2O/(TiO2)/ZnO/Al-ZnO/ITO/Glass with different ZnO thicknesses (65
nm, 130 nm, 200 nm). A fourth cell had a 10 nm thick TiO2 buffer layer between the
Cu2O and ZnO layers.
Solar Cell made with Voc (V)
Jsc (mA/cm2)
FF Efficiency
(%)
Diode ideality
factor (dark)
65 nm thick ZnO 0.19 0.64 0.39 0.05 2.7
130 nm thick ZnO 0.34 0.74 0.41 0.09 7.2
200 nm thick ZnO 0.34 0.71 0.34 0.08 7.4
65 nm thick ZnO + 10 nm thick TiO2 buffer 0.24 0.66 0.39 0.06 2.7
98
The non ideal diode characteristics of thick ZnO cells may be due to the high
resistivity of ZnO films and this will be discussed more in detail in the next section
(temperature dependent J-V analysis). The crossing of the dark and light J-V
characteristics has been reported in Cu(In,Ga)Se2 (CIGS) thin film solar cells.1,2,3
Although the underlying reasons are not clear, one of the suggested mechanisms is that
the photo-generated carriers alter the electric field distribution and change carrier
conduction during illumination.4,5,6 It seems that illumination increases the conductivity
of Cu2O and ZnO and reduces the series resistance of the solar cells. This causes the
cross-over.
To investigate the dependence of solar cell efficiency on incident light
wavelength, IPCE was measured in the wavelength range between 350 nm and 700 nm
for the solar cell with 130 nm thick ZnO and the results are shown in Figure 5.6. For the
IPCE measurement, the monochromator grating was selected and scanned to control the
wavelength of illumination light and the slit located at the exit of monochromator was
adjusted to make ~ 10 nm bandpass. The IPCE exhibits a maximum of ~ 40 % around
400 nm and abruptly decreases as wavelength increases. The IQE shows the same trend
as IPCE and this low IQE seems to limit the solar cell performance. The low IQE means
that the photo-generated carriers do not contribute to short circuit current efficiently and
the generated carriers may recombine before transport to the external load. Therefore, it
is needed to investigate the carrier transport mechanism in this Cu2O/ZnO heterojunction
solar cell.
99
20
40
60
80
100
350 400 450 500 550 600 650 7000
10
20
30
40
0
10
20
30
40LH
E (%
)
(a)
IPC
E (%
)
Wavelength (nm)
(c)
IQE
(%)
(b)
Figure 5.6. The spectrum of (a) LHE, (b) IQE and (c) IPCE of Cu2O/ZnO solar cell with
130 nm thick ZnO.
5.5 Temperature dependent J-V characterization for
Cu2O/ZnO solar cells
5.5.1 Background on the theory of temperature dependent
solar cell J-V characteristics
The current-voltage (J-V) characteristics of a p-n junction solar cell can be
interpreted within the equivalent circuit (Figure 5.2) and in the ideal case, its J-V
100
characteristics can be described by the Shockley diode equation (Equation 5.1). In the
Shockley diode model the reverse saturation current, Jo, flows by thermal generation of
minority carriers followed by diffusion to the junction and, thus, depends on the band gap
(Eg) of the semiconductor that forms the junction. For the Shockley model the Jo is given
by
Jo =qDpni
2
LpND
+ qDnni2
LnNA
= Joo exp −Eg
kT
⎛
⎝ ⎜
⎞
⎠ ⎟ (5.5)
where, ND and NA are the donor and acceptor concentrations on the n and p side of the
junction, respectively, and Dj and Li (j, i = n or p) are the diffusion coefficients and
diffusion lengths of charge carrier j, respectively. In the last term of equation (5.5), the
preexponential term, Joo, is weakly temperature dependent compared with the exponential
term. While Shockley’s model provides an approximate framework, many of the
assumptions in its derivation is often violated in real diodes. To take into account the
nonidealities, diode ideality factor A is introduced and in the forward bias region, the J-V
characteristics of p-n junction, equation (5.1), can be generalized as7
J = Jo exp qVAkT
⎛ ⎝ ⎜
⎞ ⎠ ⎟ = Joo exp − Ea
AkT⎛ ⎝ ⎜
⎞ ⎠ ⎟ exp qV
AkT⎛ ⎝ ⎜
⎞ ⎠ ⎟ (5.6)
where, the activation energy, Ea, represents the barrier that must be overcome to generate
charge carriers that contribute to the reverse saturation current and this value is equal to
the band gap in ideal diodes. In equation 5.6, forward bias approximation was assumed,
that is exp(qV/kT) >> 1. The activation energy and diode ideality factor, Ea and A,
provide insight into the phenomena occurring at the p-n junction such as tunneling and
recombination. Typically, 1 ≤ A ≤ 2 and A values larger than 2 indicate large series
101
resistances and/or the diode becoming less important in determining the J-V
characteristics of the equivalent circuit in Figure 5.2.
The charge transport and recombination mechanism of solar cells under
illumination can also be analyzed within the framework of this generalized model. The J-
V characteristics of a solar cell under illumination is given by
J = Jol exp qVAlkT
⎛
⎝ ⎜
⎞
⎠ ⎟ − Jsc = Jool exp − Ea
AlkT
⎛
⎝ ⎜
⎞
⎠ ⎟ exp qV
AlkT
⎛
⎝ ⎜
⎞
⎠ ⎟ − Jsc (5.7)
where, subscript “l” was inserted to distinguish the model parameters under illumination
from those in the dark. The parameters under illumination may be different from those in
the dark because photo-generated charge carriers may change the electric fields and
conductivities of the semiconductors, which form the junction. In equation (5.7), Jsc is
photo-generated short circuit current and a common relationship between short circuit
current and the open circuit voltage (Voc) follows the diode J-V characteristics by setting
J = 0;7,8
Voc = AlkTq
ln Jsc
Jol
⎛
⎝ ⎜
⎞
⎠ ⎟ +
Ea
q. (5.8)
Thus, if Al is independent of temperature, a plot of Voc versus T should be linear
with intercept at T = 0 K equal to the activation energy divided by q, Ea/q. If there is no
interfacial recombination and generation and if thermal generation of carriers across the
band gap dominate the saturation current then Ea will be equal Eg; otherwise, Ea will be
less than Eg. However, this method not only requires extrapolation of the data to 0 K but
assumes that Al is not a function of temperature. The diode ideality factor, Al can be
function of temperature, which introduces nonlinearities with temperature into equation
102
(5.8).9 An alternative method relies on the determination of Al and Jol at each temperature
by measuring Jsc as a function of Voc.7 Rearranging equation (5.7) and solving for ln Jsc in
terms of Voc yields
ln Jsc = qVoc
AlkT+ ln Jol . (5.9)
Thus, assuming that Al is a weak function of illumination intensity (equivalently,
a weak function of Jsc) a plot of ln Jsc versus qVoc/kT at each temperature should be linear
with slope 1/Al and intercept, ln Jol. This information can be obtained at each temperature
by collecting J-V characteristics at different illumination levels, thereby varying Jsc.
Following, activation energy can be obtained from the temperature dependence of Al and
Jol. Specifically, the relationship Jol = Jool exp(-Ea/AlkT) is rearranged to give
Al ln Jol = − Ea
kT+ Al ln Jool , (5.10)
and Ea is calculated from the slope of the straight line which results when Al ln Jol is
plotted versus 1/kT.
While this analysis provides the activation energy Ea and the temperature
dependence of the diode ideality factor, interpretation and insight into the physical origin
of these values require higher level models and a quantitative picture of the energy band
alignment and band bending at the junction. Conversely, such a picture can be
hypothesized and its predictions compared to the measurements. This is done here after
the discussion of the data analysis.
103
5.5.2. Analysis of J-V data for Cu2O/ZnO solar cells
5.5.2.1. Space Charge Limited J-V characteristics in ZnO in
the Dark
Figure 5.7 shows the dark J-V characteristics of two Cu2O/ZnO solar cells made
using a 65 nm thick ZnO film [Figure 5.7 (a)] and a 130 nm thick ZnO film [Figure 5.7
(b)] in the temperature range from 160 K to 295 K. The J-V characteristics of a third solar
cell assembled using a 200 nm thick ZnO film was similar to that shown for the solar cell
made with 130 nm thick ZnO [Figure 5.7 (b)]. If the J-V characteristics of these
Cu2O/ZnO solar cells follow Shockley’s model, the semilog plots should show a linear
region over several decades of current with slope q/AkT and 1 ≤ A ≤ 2. However, as
shown in Figure 5.7, only the solar cell made with a thin ZnO film (65 nm) shows a linear
region across at least two decades of current [Figure 5.7 (a)]. Solar cells with thicker ZnO
films [like Figure 5.7 (b)] have significant series resistance under dark and one cannot
extract information about the diode ideality factor and saturation current activation
energy from the J-V measurements.10 The nonlinear semilog J-V characteristics and the
small slope for solar cells with thick (> 65 nm) ZnO films [Figure 5.7 (b)] indicate that,
in these cases, the charge carrier transport across the solar cell and the J-V characteristics
is determined by carrier transport in the ZnO film rather than by the junction properties.
104
Figure 5.7. Dark J-V characteristics of Cu2O/ZnO solar cells as a function of temperature
between 160 K and 295 K. The thin film stack that comprised this solar cell was
Au/Cu2O/ZnO/AZO/ITO/Glass: (a) 65 nm thick ZnO (b) 130 nm thick ZnO.
(a)
(b)
105
An alternative analysis of the data in Figure 5.7 (b) is shown in Figure 5.8 (a)
wherein the J-V characteristics is plotted on a log-log scale and fitted using a power law
model rather than an exponential model. In the dark and in the low bias region (< 0.2 V),
the J-V characteristics follow a power law, J ~ Vm, where m = 1 (Ohmic behavior). At
higher bias (> 0.5 V) m > 2 and increases with decreasing temperature from ~ 5.5 at 295
K to 3 at 160 K. The power law dependence with high value of m and the temperature
dependence of m are typical of space charge limited current (SCLC)11 conduction
mechanism. The fact that power law J-V characteristics is observed when the ZnO film
thickness is increased suggests that the current is limited by the high resistivity (>107 Ω-
cm) ZnO which behaves like an insulator with a trap distribution that decays
exponentially into the forbidden band gap. Within the model developed by Rose,11 m ≈
(Tc/T) + 1 for Tc > T where kTc is the characteristic energy that describes how rapidly the
trap density distribution decays into the forbidden gap below the conduction band edge.
Figure 5.8 (b) shows a plot of m versus 1/T and from the slope of the best fit line
characteristic energy, kTc is extracted as ~ 80 meV. The intercept, however, goes through
- 0.24 instead of 1. Forcing the intercept to go through 1 and fitting the experimental data
with this constraint gives kTc = 56 meV. Either way, the conclusion is that traps
distributed several kT below the conduction band exists. The density of these traps was
estimated using the crossover method described by Kumar et al.12 In this method, the
space-charge-limited region of the temperature dependent J-V characteristics shown in
Figure 5.8 (a) are fitted to a power law and the fits are extrapolated to higher voltages
where they tend to intersect each other at a common potential called the crossover
voltage, Vc. Such an analysis is shown in Figure 5.8 (c) for the data in Figure 5.8 (a).
106
Indeed the best fit lines to data between 159 and 242 K all intersect at Vc = ~ 3 V. The
data for temperatures higher than 242 K intersect others at somewhat higher voltages but
SCLC region for these temperatures are also less developed and the slope is lower. Based
on a cross over voltage of ~ 3 V, a trap density of ~1.4 1017 cm-3 was estimated in ZnO
films.
(Figure 5.8 (a) and (b) are shown above and (c) and (d) continued on next page)
(a)
(b)
107
Figure 5.8. (a) Dark J-V characteristics (log - log scale) of a Cu2O/ZnO solar cell with
130 nm thick ZnO as a function of temperature. The solid lines are power law fits to the
data: fits for only one set of data is shown in (a). The behavior is ohmic at low bias values
with J ~ V. The exponent, m, in the power law fits depends on the temperature at high
bias values as shown in (b). (c) The “crossover” plot for the data in (a): all power law fits
are shown as lines and intersect at the crossover voltage. (d) J-V characteristics (log - log
scale) of the same solar cell in (a) but under illumination. The solid line is the power law
fit to one set of data at 295 K. Fits for other sets are not shown for clarity but the
exponent, m, in the power law fits is nearly independent of temperature in the high bias
region.
(d)
(c)
108
Figure 5.8 (d) shows the J-V characteristics of the same solar cell discussed in
Figures 5.7 (b) and 5.8 (a) on a log-log plot but under 100 mW/cm2 AM 1.5 illumination.
The J-V characteristic under illumination is very different from that measured in the dark.
In particular the power law fit in the high bias region (> 1 V) yields a temperature
independent exponent that is closer to 2 (m ~ 2.3), the value predicted by the SCLC
theory without traps.13 Although the exponent, m, is still slightly higher than 2, the J-V
characteristic under illumination seems to be determined by trap-free SCLC theory
because the photogenerated carriers fill the traps and the current flows as it would in a
trap-free material.
5.5.2.2. J-V characteristics under illumination
Under illumination, the photoconductivites of both ZnO and Cu2O are large
enough that the balance between the diode and photogenerated current dominate the J-V
characteristics at lower bias. Consequently, the diode nonideality factor Al can be
extracted from a plot of log Jsc versus Voc [Equation (5.9)]. Figure 5.9 shows such a plot
for the solar cell made with Cu2O deposited on a 130 nm thick ZnO film (the solar cell
discussed in Figure 5.8). To construct this plot, the short circuit current was varied by
changing the illumination intensity at each temperature and extracted the diode ideality
factor, Al, from the slopes ln Jsc versus qVoc/kT as a function of temperature. Indeed, a
109
plot of ln Jsc versus qVoc/kT at each temperature for all solar cells investigated in this
study were linear and resembled the typical plot shown in Figure 5.9.
Figure 5.9. Semi-log plot of short circuit current (Jsc) versus open circuit voltage (Voc) as
a function of temperature and illumination intensity for a Cu2O/ZnO heterojunction solar
cell with 130 nm thick ZnO layer. Solid lines are fits to the data using equation (5.9) to
extract Al and Jol.
110
Activation energies were extracted using both equation (5.8) and equation (5.10)
and these two methods will be referred to as the “intercept” and the “slope” methods,
respectively. In the intercept method, Ea is obtained from the T = 0 intercept of the Voc
versus T plot by extrapolation and such plots are shown for several Cu2O/ZnO solar cells
in Figure 5.10 (a). These data were collected while the solar cells were illuminated with
100 mW/cm2 AM 1.5 radiation. Although all the data from different solar cells appears
linear in the measured temperature range (100 K - 300 K), Voc begin to deviate from
linearity in the low temperature region and extrapolation to T = 0 K will introduce some
error to the extracted activation energy. In the slope method, Ea is calculated from the
slope of the line that results when Al ln Jol is plotted versus 1/T. Such plots are shown in
Figure 5.10 (b) for the same solar cells analyzed in Figure 5.10 (a). Indeed plots of Al ln
Jol versus 1/kT are linear and the slopes of the lines in Figure 5.10 (b) allow the
calculation of activation energies. The activation energies extracted by both methods are
summarized in Table 5.2. When compared, the activation energies extracted using the
intercept and slope methods are only slightly different from each other and agree to
within 10% or better. However, the slope method is more reliable because of the
uncertainty in extrapolation of Voc to T = 0 K in the intercept method.
111
Figure 5.10. (a) Open circuit voltage (Voc) as a function of temperature for solar cells
listed in Table 5.2. The lines are linear fits to the data and show extrapolation of Voc to T
= 0 K to determine the activation energy in equation (5.8) from the T = 0 K intercept. (b)
The plot of the AllnJol versus reciprocal temperature for the same solar cells as in (a). The
lines are fits to the data using equation (5.10). The activation energy for each cell is
determined from the slopes of these lines.
(a)
(b)
112
Table 5.2. Extracted saturation current activation energies for Cu2O/ZnO solar cells. The
structure of solar cells is Au/Cu2O/(TiO2)/ZnO/Al-ZnO/ITO/Glass with different ZnO
thicknesses, 65 nm, 130 nm, 200 nm and TiO2 10 nm + ZnO 65 nm. The last sample has
TiO2 buffer layer between Cu2O and ZnO layers.
Solar Cell made with Ea (eV)
from intercept Ea (eV)
from slope 65 nm thick ZnO 0.99 1.01
130 nm thick ZnO 1.21 1.26
200 nm thick ZnO 1.21 1.34
65 nm thick ZnO + 10 nm thick TiO2 buffer 1.08 1.16
5.5.3 Interpretation of the J-V-T analysis
Several significant conclusions about recombination and transport across the
Cu2O/ZnO heterojunction can be drawn from the analysis in the previous section. These
conclusions are (1) current transport across the Cu2O/ZnO junction is by recombination at
interfacial electronic states and (2) tunneling also plays a role.
5.5.3.1 Interfacial recombination
First, the activation energies extracted from the analysis in Figure 5.10 range from
0.99 to 1.34 eV, much smaller than the band gap of the Cu2O absorber layer (Eg = 2.17
eV). This unequivocally indicates that current transport across the Cu2O/ZnO interface
must involve charge recombination at interfacial electronic states in the forbidden band
gap.7,14 In this case, the measured activation energy gives clues about the Fermi levels
113
and band bending near the interface. Next, band bending at the Cu2O/ZnO junction and
an interpretation of the measured activation energy are discussed.
Since Cu2O is the absorber material, a well-designed Cu2O/ZnO heterojunction
solar cell must be asymmetrically doped to place the wider depletion region in the Cu2O
layer. Indeed the electron concentrations in the ZnO film in our solar cells are higher (~
1017 cm-3) than hole concentrations in the Cu2O layer (~ 1015 cm-3) and form an
asymmetrically doped p-n heterojunction. At such an interface, the concentration of
electrons exceed that of holes and the recombination current is limited by the supply of
holes from the Cu2O side which must overcome the barrier established by band bending
in the depletion region. The interface recombination current in this case is given by
J = qsi pi exp qVkT
⎛ ⎝ ⎜
⎞ ⎠ ⎟ (5.11)
where, si and pi are the interface recombination velocity and the hole concentration at the
interface, respectively. The hole concentration at the interface is due to holes that can
overcome the barrier established due to band bending and is given by
pi = Nv exp − qφD
kT⎛ ⎝ ⎜
⎞ ⎠ ⎟ (5.12)
where Nv is the effective density of states in the valence band and φD = (EF - Ev)/q is the
so-called diffusion potential, the difference between the Fermi level and the valence band
edge at the interface. Substitution of equation (5.12) into (5.11) yields
J = qsiNv exp − qφD
kT⎛ ⎝ ⎜
⎞ ⎠ ⎟ exp qV
kT⎛ ⎝ ⎜
⎞ ⎠ ⎟ . (5.13)
Comparison of equations (5.7) and (5.13) with Al = 1 would lead to the conclusion that
the measured activation energy, Ea is the energy difference between Fermi level and the
114
valence band edge at the interface as shown in Figure 5.11. Indeed, assuming that the
Cu2O/ZnO junction is an ideal abrupt heterojunction at thermal equilibrium and using the
known measured values of electron affinities, band gaps and carrier densities in ZnO and
Cu2O, Ea is estimated approximately ~ 1 - 1.1 eV as shown in Figure 5.11.15,16,17 This
value is approximately in the range of extracted activation energies. Thus, it is concluded
that the interface recombination is the dominant current flow mechanism across the
Cu2O/ZnO heterojunction where the rate limiting step is diffusion of holes against the
barrier established by band bending to recombine with electrons trapped in interfacial
states. This mechanism is shown as path 1 in Figure 5.11.
5.5.3.2 Tunneling
A remaining issue is that, the diode ideality factor, Al is not equal to one. A
temperature dependent value of Al indicates that tunneling may be involved in the
interfacial recombination. One possible tunneling mechanism is depicted as path 2 in
Figure 5.11 wherein a hole tunnels across the barrier due to band bending to recombine
with a trapped electron. Another clue that tunneling is involved comes from the
quantitative temperature dependence of the diode ideality factor. If the diode ideality
factor is independent of temperature then the J-V characteristic is governed by diffusion
and recombination either at the interface or in the depletion region of the junction.
However, if the diode ideality factor depends on temperature then tunneling may be
implicated. In particular, if the diode ideality factor is proportional to 1/T then the
argument of the exponent in equation (5.6) and the current become independent of
115
temperature indicating that tunneling plays a role in current flow. Thus, when tunneling
is present a plot of Al or Ad versus 1/T will be linear.
Figure 5.11. Band diagram of the Cu2O/ZnO heterojunction. EC, EV, EF, and Ea are the
conduction-band-edge, valence-band-edge, Fermi-level, and activation energies,
respectively. Most of the band bending is shown to be on the Cu2O side because of lower
doping in Cu2O compared to the ZnO side. Paths 1 and 2 indicate two possible hole
recombination pathways through interface defects. Activation energy (Ea) is the barrier
that the holes have to overcome to recombine. The approximate energy levels are based
on the properties of bulk Cu2O and ZnO. Electron affinities: 3.2 eV (Cu2O)15 and 4.2
eV16 (ZnO); band gaps: 2.17 eV17 (Cu2O) and 3.44 eV17 (ZnO); dielectric constants:
7.1117 (Cu2O) and 7.817 (ZnO), electron effective mass: 0.275mo17 (ZnO) and hole
effective mass in 0.58mo17 (Cu2O) - where mo is free electron mass. Measured carrier
concentrations of ~1015 cm-3 (Cu2O) and ~2 × 1017 cm-3 (ZnO) were used.
116
Figure 5.12 (a) shows the temperature dependence of the dark current diode
ideality factors for solar cells made with 65 nm thick ZnO film with and without a TiO2
buffer layer. The diode ideality factor shows a clear linear dependence on 1/T for the
solar cell that was made without the TiO2 film. When a thin TiO2 buffer layer is used the
diode ideality factor in the dark becomes independent of temperature at high temperatures
(low 1/T) but is still linear with 1/T at low temperatures. This indicates that
recombination via tunneling is reduced in presence of the buffer layer so that, at high
temperature, activated surface recombination becomes dominant. As temperature is
lowered, the activated recombination process slows down and tunneling becomes
important again. The two processes, recombination via tunneling and activated
recombination, can occur together and exist in parallel across the junction region.
A quantitative analytical model that attempts to describe tunneling at a junction
was introduced by Padovani and Stratton18 who were studying tunneling across a
Schottky barrier. Within this model, the diode ideality factor can be described by
A = Eoo
kTcoth Eoo
kT⎛ ⎝ ⎜
⎞ ⎠ ⎟ (5.14)
where, Eoo is a characteristic tunneling energy which depends on the doping
concentration in the semiconductor comprising a Schottky diode, Eoo=(e /2)(NA/m*ε)1/2.
Eoo is related to the tunneling probability. Thermionic field emission (tunneling) is
dominant when Eoo ~ kT/q and thermionic emission (no tunneling) is dominant when Eoo
<< kT/q.19 The line in Figure 5.12 (a) is a fit of equation (5.14) to the data for the diode
117
ideality factor of the solar cell made with 65 nm thick ZnO without the TiO2 buffer layer.
The characteristic tunneling energy, Eoo, is 54 meV. This value is difficult to explain by
invoking hole tunneling in Cu2O as shown in path 2 in Fig. 5.11 and using bulk acceptor
densities that are on the order of 1015 cm-3. Estimate of Eoo with bulk properties of Cu2O
and using acceptor densities in the 1015 cm-3 range always yields values that are much
smaller than 54 meV. Either there is a region of very high acceptor density near the
Cu2O-ZnO interfacial region or tunneling is on the ZnO side where the donor densities
are higher. This would be consistent with Zhang et al.’s proposal20 that the recombination
current in Cu2O-ZnO flows by tunneling of electrons from the ZnO conduction band into
interfacial trap states (path 3 in Figure 5.11).
Figure 5.12 (b) shows the diode ideality factors for all solar cells listed in Table 1
as a function of temperature while they are under illumination. Under illumination, the
diode ideality factors for solar cells made using 130 nm and 200 nm thick ZnO layers are
independent of temperature indicating that tunneling is absent. Tunneling is still present
in the solar cell made using thinner (65 nm) ZnO film. That presence of tunneling
depends on the ZnO film thickness supports the previous conclusion that the tunneling is
on the ZnO side. That the tunneling is eliminated under illumination is consistent with the
proposal that the tunneling is into interfacial traps: under illumination interfacial traps are
filled which would eliminate tunneling.
118
Figure 5.12. (a) Diode ideality factor versus reciprocal temperature for J-V
characteristics acquired in the dark from solar cells made using 65 nm thick ZnO with
and without a 10 nm thick TiO2 buffer layer. The solid line for the solar cell without the
buffer layer is a fit to equation (5.16). The solid line for the solar cell with the buffer
layer is drawn to guide the eye. (b) Diode ideality factor versus reciprocal temperature for
J-V characteristics acquired under 100 mW/cm2 illumination from solar cells listed in
Table 5.2.
(a)
(b)
1000/T (K-1)
1000/T (K-1)
119
5.5.3.3 Auger depth profile
It is well known that copper can diffuse readily even at low temperature. To
evaluate possible metal interdiffusion at the ITO-ZnO and the Cu2O-ZnO interfaces, the
variation of Cu, Zn, and In throughout the solar cell film stack was measured using depth
profiling by Auger spectroscopy. In this film the Cu2O and ZnO film thicknesses were
280 nm and 65 nm respectively. Figure 5.13 shows the Cu, Zn and In depth profiles. The
sputtering rate is material and composition dependent and no attempt was made to
calibrate the sputtering rate. From the approximate locations of the interfaces and ex situ
measurements of the film thickness, it was estimated that each sputtering cycle removes ~
2.5 nm/cycle in Cu2O and ~ 1.3 nm/cycle in ZnO. Clearly, there is significant
interdiffusion of Zn and Cu at the Cu2O-ZnO interface. The initial Cu2O-ZnO interface
seems to be where the maximum in the Zn and minimum in the Cu concentration occur.
There could be preferential segregation at this interface and abundance of Zn due to the
lattice mismatch and grain boundaries. While it may seem that the Cu is diffusing up a
concentration gradient into ZnO this is likely due to differences in the partition
coefficient for Cu in ZnO and Cu2O, i.e., chemical potential gradients drive the diffusion
of Cu into ZnO. Clearly, Cu diffuses all the way into ZnO and even ITO. The width of
the transition region is ~ 25 cycles which corresponds to a thickness of ~50 nm much
thicker than the roughness of the interface. Thus, the observed gradients are clearly due to
diffusion and not due to interfacial roughness. Copper can be a p-type dopant and
diffusion of Cu into ZnO can compensate the n-type ZnO film. On the other hand Zn is
an n-type dopant in Cu2O and can compensate the p-type film. This interdiffusion and
compensation may be responsible for the low open circuit voltages observed in Cu2O-
120
ZnO heterojunction solar cells.
Figure 5.13. Copper, zinc and indium concentration profile measured using Auger depth
profiling.
121
References
1 A. Rothwarf, Sol. Cells 2, 115 (1980).
2 A. Kylner, J. Appl. Phys. 85, 6858 (1999).
3 G. Friesen, E. D. Dunlop, R. Wendt, Thin Solid Films 387, 239 (2001).
4 K. Leschkies, T. J. Beatty, M. S. Kang, D. Norris, E. S. Aydil, ACS Nano 3, 3638
(2009).
5 C. G. Walle, J. Phys. Condens. Matter. 20, 064230 (2008).
6 R. R. Potter, J. R. Sites, IEEE T. Electron Dev. 31, 571 (1984).
7 V. Nadenau, U. Rau, A. Jasenek, H. W. Schock, J. Appl. Phys. 87, 584 (2000).
8 M. Turcu, O. Pakma, U. Rau, Appl. Phys. Lett. 80, 2598 (2002).
9 S. S. Hegedus, J. Appl. Phys. 63, 5126 (1988).
10 M. -K. Han, P. Sung, W. A. Anderson, IEEE Elect. Dev. Lett. EDL-3, 121 (1982).
11 A. Rose, Phys. Rev. 97, 1538 (1955).
12 V. Kumar, S. C. Jain, A. K. Kapoor, J. Poortmans, R. Mertens, J. Appl. Phys. 94, 1283
(2003).
13 P. Mark, W. Helfrich, J. Appl. Phys. 33, 205 (1962).
14 Richard H. Bube, Photoelectronic Properties of Semiconductors (Cambridge
University Press, Cambridge, 1992) pp. 256.
15 S. S. Jeong, A. Mittiga, E. Salza, A. Masci, S. Passerini, Electrochim. Acta 53, 2226
(2008).
16 H. Kobayashi, H. Mori, T. Ishida, Y. Nakato, J. Appl. Phys., 77, 1301 (1995).
17 Otfried Madelung, Semiconductors : Data Handbook (Springer, 3rd edition, 2004).
122
18 F. A. Padovani, R. Stratton, Solid State Electronics, 9, 695 (1966)
19 A. Gumus, A. Turut, N. Yalcin, J. Appl. Phys., 91, 254 (2002)
20 D. K. Zhang, Y. C. Liu, Y. L. Liu, H. Yang, Physica B, 351, 178-183 (2004).
123
Chapter 6
Summary and future direction
6.1 Summary
Among the metal oxides which exhibit a variety of functional properties
depending on their crystal structure and bonding between the metal cation and oxygen, I
focused on zinc oxide (ZnO) and cuprous oxide (Cu2O). ZnO is always an n-type
semiconductor either due to oxygen vacancies1 or unintentional hydrogen incorporation,2
and its wide band gap (~ 3.4 eV) makes it a suitable window layer in solar cells. On the
other hand, Cu2O always shows p-type conductivity, which is usually attributed to Cu
vacancies.3 The band gap of Cu2O (~ 2.2 eV), its high absorption coefficient4 and long
minority carrier diffusion length,5 make it an attractive absorber material in solar cells. In
addition to their optical and electrical properties, the facts that they are benign
environmentally, abundant on earth and cheap motivated this research to apply these
materials for thin film solar cell applications.
The main topics of my research are (1) the properties of ZnO thin films deposited
by metal organic chemical vapor deposition (MOCVD), plasma enhanced CVD
(PECVD) and sputtering, (2) the properties of Cu2O thin films deposited by MOCVD and
(3) Cu2O/ZnO heterojunction and solar cell applications.
124
ZnO thin films deposited by MOCVD from zinc acetylacetonate (Zn(C5H7O2)2,
Zn(acac)2) shows textured structure with [0002]ZnO parallel to the substrate normal on
various substrates including amorphous glass, silicon, sapphire (a-plane). The film
morphology is usually columnar with rough surface because of the anisotropic growth
rate of ZnO [ (0001) (1011) (1010)> > ]. However, the ZnO films deposited on r-plane
sapphire show epitaxy with 2 3ZnO Al O(1120) (1102) and
2 3ZnO Al O[0001] [1120]⊥ and the
surface is smooth when deposited at low temperature (below 500˚C). X-ray diffraction
(XRD) results show that the films deposited at high temperature (above 600˚C) have high
quality (small full width half maximum (FWHM) of θ-2θ scan and rocking curve) but
have poles grow out of surface. The films deposited at low temperature have stress to
maintain the epitaxy but this stress is relieved by growing poles at high deposition
temperature. These poles are suppressed by two step temperature deposition such that the
films are grown at low temperature at first and then continued at high temperature.
ZnO films deposited by magnetron sputtering with undoped ZnO target have
same textured structure ([0002]ZnO is parallel to substrate normal) as MOCVD. The grain
size increases with oxygen partial pressure increase but the deposition rate reduces. Post
annealing treatments such as thermal annealing and oxygen plasma exposure increases
grain size and relieve film stress and this annealing effect is prominent in thin film case
than thick film. From the investigation of the crystal structure dependence on annealing
temperature, it was found that the film stress is not relieved below 150˚C and the
annealing effect increases rapidly with temperature and saturated above 350˚C. The grain
sizes also are not affected by annealing temperature below 150˚C and then increase
125
monotonously along with annealing temperature. The RMS roughness of films deposited
at room temperature is ~ 3 nm and oxygen plasma has increases the roughness to ~ 8 nm.
The electrical resistivity of films deposited at room temperature is above the limit of the
measurement unit (~ 107 Ω-cm) and high deposition temperature (350˚C) reduces the
resistivity by ~ 4 orders.
Cu2O thin films were deposited on ZnO coated glass substrates by MOCVD using
copper(II) hexafluoroacetylacetonate [Cu(C5HF6O2)2], oxygen gas and water vapor. The
Cu2O films are single phase with optical band gap, around 2 eV. XRD shows that Cu2O
thin films grow on ZnO with preferred (220)Cu2O || (0002)ZnO orientation, textured
structure on glass substrate. The grain size and stress of Cu2O films increase with
substrate temperature, and this structural property change results in the increase of
electrical mobility. The property dependence on film thickness shows that grain size and
film stress are inversely proportional to film thickness and as a result, mobility decrease.
The Cu2O films also has in-plane rotational alignment with the epitaxial relationship,
(220)Cu2O || (0002)ZnO ; [001]Cu2O || [1210]ZnO. Coincidence lattice model is invoked to
explain this epitaxial relation.
Cu2O/ZnO heterojunction thin film solar cells were fabricated based on MOCVD
deposited Cu2O and sputtered ZnO. The carrier transport and recombination mechanisms
at the heterojunction were investigated by analyzing the temperature dependent J-V
characteristics under dark and varying levels of illumination. The saturation current
activation energies varied between 0.99 eV to 1.34 eV depending on how the solar cell
was prepared. The activation energies are smaller than the band gap of Cu2O (2.2 eV)
and this reveals that interface recombination is the dominant carrier transport mechanism
126
in these solar cells. Tunneling plays an important role in interface recombination and
tunneling characteristic energy is 54 meV. A buffer layer at the interface can reduce
tunneling and improvements in the performance of Cu2O/ZnO heterojunction solar cells
will require suppression recombination and tunneling at the interface.
6.2 Future direction
The analysis of carrier transport mechanism through Cu2O/ZnO heterojunction
suggests that the suppression interface recombination and tunneling will improve the
solar cell performance. Thin TiO2 buffer layer shows an effect on reducing tunneling.
Buffer layer at the junction interface can change the transport mechanism by passivating
interface defects and modifying band alignment therefore, finding effective buffer layer
in Cu2O/ZnO heterojunction is important. The buffer layer should have appropriate
electronic properties to make favorable band alignment and in this point of view, zinc
sulfide (ZnS) can be a good candidate. Figure 6.1 shows the band diagram of Cu2O, ZnS
and ZnO. The smaller electron affinity and the larger band gap of ZnS than ZnO have a
possibility to produce larger build-in potential at the interface between Cu2O and ZnS
because ZnS may have larger fermi-level difference with Cu2O than ZnO. ZnS does not
make any additional absorption loss of incident light compared to solar cells without ZnS.
Because of the larger band gap of ZnS than ZnO, the illuminated light through ZnO layer
will not be absorbed in ZnS layer. In addition to the electronic and optical properties, ZnS
is also non toxic and cheap material therefore, it is a promising buffer layer for
Cu2O/ZnO heterojunction solar cell.
127
Figure 6.1. Band gap diagram of Cu2O, ZnS and ZnO. Ec, Ef and Ev are the conduction
band edge, fermi-level and valence band edge, respectively. The dopant binding energy is
arbitrarily selected therefore the positions of Ef are arbitrary. The electron affinity and
band gap of each materials are come from the references 6,7,8,9.
Solar cell performance can be improved by making use of nano-structure. The use
of nano-structure provides several advantages for solar cells. First, nano-structure can
improve light harvesting efficiency (LHE) because light scattering can be occurred within
nano-structure and this may enhance the duration time of light in the material, hence light
absorption can be increased. Recently, the improvement of light absorption in solar cells
utilizing nano-structures (nano-wire, nano-dome etc.) is reported by several groups.10,11,12
Second, carrier transport can be improved because nano-structures have large surface
area compared to flat film which can enlarge junction interface areas. If the
photogenerated current density is same, the nano-structured solar cell will produce larger
128
total current than flat film solar cell. Nano-structured solar cell based on Cu2O/ZnO
heterojunction is reported by a few groups.13,14 Hsueh et. al. reported a nano-wire solar
cell based on Cu2O/ZnO heterojunction in which Cu2O was deposited on ZnO nano-
wires. Although they tried to make use of nano-wire, the deposited Cu2O did not
penetrate between the nano-wires and just stacked on top of the wire ends therefore Cu2O
film contacts ZnO only at the end of ZnO wire tips. In this structure, the cell cannot seem
to make use of the full advantage of nano-structure. This may be due to the fact that they
used sputtering to deposit Cu2O on ZnO wire. The test deposition of Cu2O by MOCVD
on ZnO nano-wire shows that the deposited Cu2O penetrated between wires as shown in
Figure 6.2. Therefore, nano-structured Cu2O/ZnO solar cell by CVD is a potential
subject.
Figure 6.2. Cross sectional SEM image of ZnO nanowires grown on ITO coated glass
substrate with and without Cu2O film deposited by MOCVD. (a) is the ZnO nanowires
grown on ITO glass before the Cu2O deposition. (b) is the after the Cu2O deposition.
(a) (b)
ZnO wire
ITO Glass 500 nm
ZnO wire + Cu2O
ITO/Glass1 µm
129
References
1 D. C. Look, J. W. Hemsky, J. R. Sizelove, Phys. Rev. Lett. 82, 2552 (1999).
2 Chris G. Van de Walle, Phy. Rev. Lett. 85, 1012 (2000).
3 M. Nolan, S. D. Elliott, Phys. Chem. Chem. Phys. 8, 5350 (2006).
4 A. E. Rakhshani, Solid-State Electronics, 29, 7 (1986).
5 C. A. Dimitriadis, L. Papadimitriou, N. A. Economou, J. Mater. Sci. Lett. 2, 691 (1983).
6 S. S. Jeong, A. Mittiga, E. Salza, A. Masci, S. Passerini, Electrochimica Acta 53,226
(2008).
7 H. Kobayashi, H. Mori, T. Ishida, Y. Nakato, J. Appl. Phys., 77, 1301 (1995).
8 Otfried Madelung, Semiconductors : Data Handbook (Springer, 3rd edition, 2004).
9 M. M. Islam, S. Ishizuka, A. Yamada, K. Sakurai, S. Niki, T. Sakurai, K. Akimoto, Sol.
Energ. Mat. Sol. C. 93, 970 (2009).
10 M. D. Kelzenberg, S. W. Boettcher, J. A. Petykiewicz, D. B. Turner-Evans, M. C.
Putnam, E. L. Warren, J. M. Spurgeon, R. M. Briggs, N. S. Lewis, H. A. Atwater, Nat.
Mater. 9, 239 (2010).
11 C. H. Peters, A. R. Guichard, A. C. Hryciw, M. L. Brongersma, M. D. McGehee, J.
Appl. Phys. 105, 124509 (2009).
12 J. Zhu, C. Hsu, Z. Yu, S. Fan, Y. Cui, Nano lett. 10, 1979 (2010).
13 T. J. Hsueh, C. L. Hsu, S. J. Chang, P. W. Guo, J. H. Hsieh, I. C. Chen, Scripta Mater.
57, 53 (2007).
14 B. D. Yuhas, P. Yang, J. Am. Chem. Soc. 131, 3756 (2009).
130
Bibliography
A. E. Rakhshani, Solid State Electron. 29, 7 (1986).
A. Gumus, A. Turut, N. Yalcin, J. Appl. Phys. 91, 254 (2002).
A. Kylner, J. Appl. Phys. 85, 6858 (1999).
A. Mittiga, E. Salza, F. Sarto, M. Tucci, R. Vasanthi, Appl. Phys. Lett. 88, 163502 (2006).
A. Ohtomo, A. Tsukazaki, Semicond. Sci. Tech. 20, S1 (2005).
A. O. Musa, T. Akomolafe, M. J. Carter, Sol. Energ. Mat. Sol. C. 51, 305 (1998).
A. Rose, Phys. Rev. 97, 1538 (1955).
A. Trampert K. H. Ploog, Cryst. Res. Technol. 35, 793 (2000). A. Rothwarf, Sol. Cells 2,
115 (1980).
B. Angadi, H. C. Park, H. W. Choi, J. W. Choi, W. K. Choi, J. Phys. D Appl. Phys. 40,
1422 (2007).
B. Claflin, D. C. Look, S. J. Park, G. Cantwell, J. Cryst. Growth 287, 16 (2006).
B. Cao, W. Cai, H. Zeng, G. Duan, J. Appl. Phys. 99, 73516 (2006).
B. D. Yuhas, P. Yang, J. Am. Chem. Soc. 131, 3756 (2009).
B. P. Rai, Sol. Cells 25, 265 (1988).
B. P. Zhang, N. T. Binh, K. Wakatsuki, N. Usami, Y. Segawa, Appl. Phys. A 78, 25
(2004).
B. S. Li, Y. C. Liu, Z. S. Chu, D. Z. Shen, Y. M. Lu, J. Y. Zhang, X. W. Fan, J. Appl.
Phys. 91, 501 (2002).
B. S. Li, Y. C. Liu, D. Z. Shen, Y. M. Lu, J. Y. Zhang, X. G. Kong, X. W. Fan, Z. Z. Zhi,
J. Vac. Sci. Technol. A 20, 265 (2002).
131
C. A. Dimitriadis, L. Papadimitriou, N. A. Economou, J. Mat. Sci. Lett. 2, 691 (1983).
C. A. Wolden, T. M. Bames, J. B. Baxter, E. S. Aydil, J. Appl. Phys. 97, 43522 (2005).
Chris G. Van de Walle, Phy. Rev. Lett. 85, 1012 (2000).
C. G. Walle, J. Phys. Condens. Matter. 20, 64230 (2008).
C. H. Bates, W. B. White, R. Roy, Science 137, 993 (1962).
C. H. Peters, A. R. Guichard, A. C. Hryciw, M. L. Brongersma, M. D. McGehee, J. Appl.
Phys. 105, 124509 (2009).
C. N. R. Rao, B. Raveau, Transition Metal Oxides, Wiley-VCH, New York, 1998.
D. C. Kim, B. H. Kong, H. K. Cho, D. J. Park, J. Y. Lee, Nanotechnology 18, 15603
(2007).
D. C. Look, Mat. Sci. Eng. B 80, 383 (2001).
D. C. Look, J. W. Hemsky, J. R. Sizelove, Phys. Rev. Lett. 82, 2552 (1999).
D. K. Zhang, Y. C. Liu, Y. L. Liu, H. Yang, Physica B, 351, 178-183 (2004).
David P. Norton, Mater. Sci. Eng. R 43, 139 (2004).
E. Mirica, G. Kowach, P. Evans, H. Du, Cryst. Growth Des. 4, 147 (2004).
F. A. Padovani, R. Stratton, Solid State Electronics 9, 695 (1966).
G. Friesen, E. D. Dunlop, R. Wendt, Thin Solid Films 387, 239 (2001).
G. G. condorelli, G. Malandrino, I. Fragala, Chem. Mater. 6, 1861 (1994).
G. P. Pollack, D. Trivich, J. Appl. Phys. 46, 163 (1975).
G. Xiong, J. Wilkinson, B. Mischuck, S. Tuzemen, K. B. Ucer, R. T. Williams, Appl.
Phys. Lett. 80, 1195 (2002).
H. Kobayashi, H. Mori, T. Ishida, Y. Nakato, J. Appl. Phys. 77, 1301 (1995).
132
H. Karzel, W. Potzel, M. Köfferlein, W. Schiessl, M. Steiner, U. Hiller, G. M. Kalvius, D.
W. Mitchell, T. P. Das, P. Blaha, K. Schwarz, M. P. Pasternak, Phys. Rev. B 53, 11425
(1996).
H. Raebiger, S. Lany, A. Zunger, Phys. Rev. B 76, 45209 (2007).
H. Tanaka, T. Shimakawa, T. Miyata, H. Sato, T. Minami, Thin Solid Films 469-470, 80
(2004).
I. Volintiru, M. Creatore, B. J. Kniknie, C. I. M. A. Spee, M. C. M. Sanden, J. Appl. Phys.
102, 43709 (2007).
J. B. Baxter, E. S. Aydil, J. Cryst. Growth 274, 407 (2005).
J. B. Mooney, S. B. Radding, Annu. Rev. Mater. Sci. 12, 81 (1982).
J. F. Wager, Science 300, 1245 (2003).
J. Herion, E. A. Niekisch, G. Scharl, Sol. Energ. Mater. 4, 101 (1980).
J. H. Park, S. J. Jang, S. S. Kim, B. T. Lee, Appl. Phys. Lett. 89, 121108 (2006).
J. J. Loferski, J. Appl. Phys. 27, 777 (1956).
J. Q. Xie, J. W. Dong, A. Osinsky, P. P. Chow, T. W. Heo, D. P. Norton, S. J. Pearton, X.
Y. Dong, C. Adelmann, C. J. Palmstrom, Mater. Res. Soc. Symp. P. 891, 407 (2006).
J. Ramirez-Ortiz, T. Ogurab, J. Medina-Valtierra, S. E. Acosta-Ortiz, P. Bosch, J. A. de
los Reyes, V. H. Lara, Appl Surf. Sci. 174, 177-184 (2001).
J. Zhu, C. Hsu, Z. Yu, S. Fan, Y. Cui, Nano lett. 10, 1979 (2010).
K. Akimoto, S. Ishizuka, M. Yanagita, Y. Nawa, G. K. Paul, T. Sakurai, Sol. Energy 80,
715 (2006).
K. Ellmer, G. Vollweiler, Thin Solid Films 496, 104 (2006).
K. Ellmer, J. Phys. D Appl. Phys. 33, R17 (2000).
133
K. Ellmer, J. Phys. D. Appl. Phys. 34, 3097 (2001).
K. Koike, I. Nakashima, K. Hashimoto, S. Sasa, M. Inoue, M. Yano, Appl. Phys. Lett. 87,
112106 (2005).
K. Leschkies, T. J. Beatty, M. S. Kang, D. Norris, E. S. Aydil, ACS Nano 3, 3638 (2009).
L. C. Olsen, F. W. Addis, W. Miller, Sol. Cells 7, 247 (1982-1983).
L. Papadimitriou, N. A. Economou D. Trivich, Sol. Cells 3, 73 (1981).
L. T. Drapak, Semiconductors 2, 624 (1968).
M. Chen, Z. L. Pei, C. Sun, J. Gong, R. F. Huang, L. S. Wen, Mater. Sci. Eng. B 85, 212
(2001).
M.Chen, Z. L. Pei, X. Wang, C. Sun, L. S. Wen, J. Vac. Sci. Technol. A 19, 963 (2001).
M. Chen, Z. L. Pei, X. Wang, Y. H. Yu, C. Sun, L. S. Wen, J. Phys. D Appl. Phys. 33,
2538 (2000).
M. D. Kelzenberg, S. W. Boettcher, J. A. Petykiewicz, D. B. Turner-Evans, M. C.
Putnam, E. L. Warren, J. M. Spurgeon, R. M. Briggs, N. S. Lewis, H. A. Atwater, Nat.
Mater. 9, 239 (2010).
M. Izaki, T. Shinagawa, K. Mizuno, Y. Ida, M. Inaba, A. Tasaka, J. Phys. D Appl. Phys.
40, 3326 (2007).
M. K. Han, P. Sung, W. A. Anderson, IEEE Elect. Dev. Lett. EDL-3, 121 (1982).
M. Kim, Y. Hong, J. Yoo, G. Yi, G. Park, K. Kong, H. Chang, Phys. Stat. Sol. (RR) 2,
197 (2008).
M. Liu, H. K. Kim, Appl. Phys. Lett. 84, 173 (2004).
M. M. Islam, S. Ishizuka, A. Yamada, K. Sakurai, S. Niki, T. Sakurai, K. Akimoto, Sol.
Energ. Mat. Sol. C. 93, 970 (2009).
134
M. Nolan, S. D. Elliott, Phys. Chem. Chem. Phys. 8, 5350 (2006).
M. Ohring, Materials Science of thin films, Academic press, USA, 2002.
M. Turcu, O. Pakma, U. Rau, Appl. Phys. Lett. 80, 2598 (2002).
N. Tsuji, H. Komiyama, K. Tanaka, J. J. Appl. Phys. 29, 835 (1990).
Otfried Madelung, Semiconductors : Data Handbook, Springer, 3rd edition, 2004.
P. Kung, C. J. Sun, A. Saxler, H. Ohsato, M. Razeghi, J. Appl. Phys. 75, 4515 (1994).
P. Mark, W. Helfrich, J. Appl. Phys. 33, 205 (1962).
Q. Ma, Z. Ye, H. He, L. Zhu, J. Wang, B. Zhao, Mater. Lett. 61, 2460 (2007).
R. G. Gordon, MRS Bull. 52, August (2000).
R. G. Gordon, Mater. Res. Soc. Symp. P. 426, 419 (1996).
Richard H. Bube, Photoelectronic Properties of Semiconductors (Cambridge University
Press, Cambridge, 1992) pp. 256.
R. L. Hoffman, B. J. Norris, J. F. Wager, Appl. Phys. Lett. 82, 733 (2003).
R. R. Potter, J. R. Sites, IEEE T. Electron Dev. 31, 571 (1984).
R. Swanepoel, J. Phys. E Sci. Instrum. 16, 1214 (1983).
R. Triboulet, 10th European Workshop on MOVPE, Lecc (Italy) 11 Aug June (2003).
R. T. Downs, K. L. Bartelmehs, G. V. Gibbs, Am. Mineral. 78, 1104 (1993).
S. Brehme, F. Fenske, W. Fuhs, E. Nebauer, M. Poschenrieder, B. Selle, I. Sieber, Thin
Solid Films 342, 167 (1999).
S. B. Krupanidhi, M. Sayer, J. Appl. Phys. 56, 3308 (1984).
Satischandra B. Ogale, Thin films and Heterostructures for oxide electronics, Chapter 10,
303 (Springer, 2005).
135
S. Ishizuka, K. Suzuki, Y. Okamoto, M. Yanagita, T. Sakurai, K. Akimoto, N. Fujiwara,
H. Kobayashi, K. Matsubara, S. Niki, Phys. Status. Solidi. 1, 1067 (2004).
S. Ishizuka, S. Kato, Y. Okamoto, K. Akimoto, Appl. Phys. Lett. 80, 950 (2002).
S. Ishizuka, S. Kato, Y. Okamoto, K. Akimoto, J. Cryst. Growth 237, 616 (2002).
S. J. Pearton, W. H. Heo, M. Ivill, D. P. Norton, T. Steiner, Semicond. Sci. Tech. 19,
R59-R74 (2004).
S. S. Hegedus, J. Appl. Phys. 63, 5126 (1988).
S. S. Jeong, A. Mittiga, E. Salza, A. Masci, S. Passerni, Electrochim. Acta. 53, 2226
(2008).
T. ITO, H. Yamaguchi, K. Okabe, T. Masumi, J. Mater. Sci. 33, 3555 (1998).
T. J. Hsueh, C. L. Hsu, S. J. Chang, P. W. Guo, J. H. Hsieh, I. C. Chen, Scripta Mater. 57,
53 (2007).
T. Minami, MRS Bull. 38, August (2000).
T. Minami, T. Miyata, K. Ihara, Y. Minamino, S. Tsukata, Thin Solid Films 494, 47
(2006).
U. Ozgur, Ya. I. Alivov, C. Liu, A, Teke, M. A. Reshchikov, S. Dogan, V. Avrutin, S. J.
Cho, H. Morkoc, J. Appl. Phys. 98, 41301 (2005).
V. Kumar, S. C. Jain, A. K. Kapoor, J. Poortmans, R. Mertens, J. Appl. Phys. 94, 1283
(2003).
V. Nadenau, U. Rau, A. Jasenek, H. W. Schock, J. Appl. Phys. 87, 584 (2000).
V. Tvarozek, K. Shtereva, I. Novotny, J. Kovac, P. Sutta, R. Srnanek, A. Vincze,
Vacuum 82, 166 (2008).
W. Shockley, H. J. Queisser, J. Appl. Phys. 32, 510 (1961).
136
W. Y. Ching, Y. N. Xu, K. W. Wong, Phys. Rev. B 40, 7684 (1989).
Y. Chen, D. M. Bagnall, H. Koh, K. Park, K. Hiraga, Z. Zhu, T. Yao, J. Appl. Phys. 84,
3912 (1998).
Y. J. Kim, H. J. Kim, Mater. Lett. 21, 351 (1994).
Y. Nakano, S. Saeki, T. Morikawa, Appl. Phys. Lett. 94, 22111 (2009).
Y. W. Heo, S. J. Park, K. Ip, S. J. Pearton, D. P. Norton, Appl. Phys. Lett. 83, 1128
(2003).
Z. Li, W. Gao, Mater. Lett. 58, 1363 (2004).